[pve-devel] [PATCH-SERIES qemu/common/storage/qemu-server/container/manager v6 00/37] backup provider API

[pve-devel] [PATCH-SERIES qemu/common/storage/qemu-server/container/manager v6 00/37] backup provider API

[Fiona Ebner]

v5: https://lore.proxmox.com/pve-devel/20250321134852.103871-1-f.ebner@proxmox.com/
v4: https://lore.proxmox.com/pve-devel/20241114150754.374376-1-f.ebner@proxmox.com/
v3: https://lore.proxmox.com/pve-devel/20241107165146.125935-1-f.ebner@proxmox.com/

Changes in v6:
* Factor out some helpers in QEMU backup code.
* Import constants in LXC::Namespace module from PVE::Tools instead of
  re-declaring.
* Require that source for 'qemu-img' restore mechanisms is in raw
  format.
* Document that methods for extracting configurations are called
  without restore_*_init() first.
* Rename restore_get_*_config() methods to archive_get_*_config().
* Replace job/backup hooks with dedicated methods,
  {job,backup}_{init,cleanup} and backup_container_prepare() for the
  privileged part right before backup_container() is called.
* Remove backup_get_archive_name() method. The archive name is now
  returned as part of the backup_init() method. That was also a bit of
  a misnomer, since it was necessary for the backup provider to
  remember that archive name for later.
* Supersede backup_get_task_size() method and return stats as part of
  the backup_cleanup() method (in the success case). Currently, this
  only includes the archive size as well, but could be extended later
  to allow custom stats to include in the task notification summary.
* Drop reference to vm_backup_get_device_info() method in docs/cover
  letter. This was dropped in an earlier version.
* Squash implementation of external backup in qemu-server together
  with patch with untrusted check.

Changes in v5:
* Drop already applied patches.
* Rebase on latest master.
* Set finishing state and end time in backup state in QEMU to avoid
  wrongly detecting previous backup as still active leading to a
  warning.
* Unbreak container restore with a bandwidth limit.
* Switch from 'block-device' mechanism to 'file-handle', providing
  access to the image contents as a (virtual) file created with
  'nbdfuse'. The reason is that block devices created via qemu-nbd
  will lead to a lot of ugly error messages if the device is
  disconnected before partition probing is finished (which can be
  delayed even until the end of the backup, apparently there is a
  lock). This also avoids the need to load the 'nbd' kernel module
  and not exposing the devices as host block devices is just nicer,
  potentially also security-wise. This requires using the fallocate
  syscall instead of the previous BLKDISCARD ioctl. A helper for that
  is provided via the Storage/Common.pm module.
* Indicate support for QEMU feature via 'backup-access-api' flag
  returned by 'query-proxmox-support' QMP command rather than
  requiring a version guard, so it can land whenever.
* Update storage's ApiChangelog, split out API age+version bump into
  its own patch.
* Let plugins define their own list of sensitive properties.
* Add postinst snippet to load NBD module during qemu-server upgrade.
* Check nbd/parameters directory instead of nbd/coresize to determine
  whether the module is loaded. coresize was just picked on a whim,
  parameters seems cleaner.
* Handle edge case with size not being a single number for character
  devices when parsing/checking tar contents.
* Borg plugin:
  * improve SSH options/handling
  * improve handling and permissions of run and ssh directories
  * mount borg archives to subdirectory of run dir
  * add helper for common command environment
  * move to Custom subfolder
* Directory example plugin:
  * use NBD device node reservation mechanism.
  * die if NBD module is not loaded.
  * Unbreak logging warnings by using the correct function.

Changes in v4:
* Drop already applied patches.
* Rework run_in_userns() and related helpers.
* Improve child PID handling for VM 'block-device' method backup
* Improve child PID and volume cleanup for VM 'block-device' method
  backup
* Improve /dev/nbdX handling by adding reservation and checks
* Add modules-load and modprobe configs for 'nbd' module
* Add some more logging

Changes in v3:
* Add storage_has_feature() helper and use it to decide on whether the
  storage uses a backup provider, instead of having this be implicit
  with whether a backup provider is returned by new_backup_provider().
* Fix querying block-node size for fleecing in stop mode, by issuing
  the QMP command only after the VM is enforced running.
* Run backup_container() in user namespace associated to the
  container.
* And introduce 'prepare' phase for backup_hook() to be used to
  prepare for running in that user namespace context.
* Pass in guest and firewall config as raw data instead of by file
  name (so files don't have to be accessible in user namespace context
  for containers).
* Run restore of containers with 'directory' mechanism in user
  namespace switching from 'rsync' to 'tar' which is easier to "split"
  into a privileged and unprivileged half.
* Check potentially untrusted tar archives.
* Borg plugin: make SSH work and use that.

Changes in v2:
* Add 'block-device' backup mechansim for VMs. The NBD export is
  mounted by Proxmox VE and only the block device path (as well as a
  callback to get the next dirty range for bitmaps) is passed to the
  backup provider.
* Add POC example for Borg - note that I tested with borg 1.2.4 in
  Debian and only tested with a local repository, not SSH yet.
* Merge hook API into a single function for backup and for jobs.
* Add restore_vm_init() and restore_vm_cleanup() for better
  flexibility to allow preparing the whole restore. Question is
  if restore_vm_volume_init() and restore_vm_volume_cleanup() should
  be dropped (but certain providers might prefer using only those)?
  Having both is more flexible, but makes the API longer of course.
* Switch to backup_vm() (was per-volume backup_vm_volume() before) and
  backup_container(), passing along the configuration files, rather
  than having dedicated methods for the configuration files, for
  giving the backup provider more flexibility.
* Some renames in API methods/params to improve clarity.
* Pass backup time to backup 'start' hook and use that in the
  directory example rather than the job start time.
* Use POD for base plugin documentation and flesh out documentation.
* Use 'BackupProvider::Plugin::' namespace.
* Various smaller improvements in the directory provider example.

======

A backup provider needs to implement a storage plugin as well as a
backup provider plugin. The storage plugin is for integration in
Proxmox VE's front-end, so users can manage the backups via
UI/API/CLI. The backup provider plugin is for interfacing with the
backup provider's backend to integrate backup and restore with that
backend into Proxmox VE.

This is an initial draft of an API and required changes to the backup
stack in Proxmox VE to make it work. Depending on feedback from other
developers and interested parties, it can still substantially change.

======

The backup provider API is split into two parts, both of which again
need different implementations for VM and LXC guests:

1. Backup API

There are two hook callback functions, namely:
1. job_hook() is called during the start/end/abort phases of the whole
   backup job.
2. backup_hook() is called during the start/end/abort phases of the
   backup of an individual guest. There also is a 'prepare' phase
   useful for container backups, because the backup method for
   containers itself is executed in the user namespace context
   associated to the container.

The backup_get_mechanism() method is used to decide on the backup
mechanism. Currently, 'file-handle' or 'nbd' for VMs, and 'directory'
for containers is possible. The method also let's the plugin indicate
whether to use a bitmap for incremental VM backup or not. It is enough
to implement one mechanism for VMs and one mechanism for containers.

Next, there are methods for backing up the guest's configuration and
data, backup_vm() for VM backup and backup_container() for container
backup, with the latter running

Finally, some helpers like getting the provider name or volume ID for
the backup target, as well as for handling the backup log.

1.1 Backup Mechanisms

VM:

Access to the data on the VM's disk from the time the backup started
is made available via a so-called "snapshot access". This is either
the full image, or in case a bitmap is used, the dirty parts of the
image since the last time the bitmap was used for a successful backup.
Reading outside of the dirty parts will result in an error. After
backing up each part of the disk, it should be discarded in the export
to avoid unnecessary space usage on the Proxmox VE side (there is an
associated fleecing image).

VM mechanism 'file-handle':

The snapshot access is exposed via a file descriptor. A subroutine to
read the dirty regions for incremental backup is provided as well.

VM mechanism 'nbd':

The snapshot access and, if used, bitmap are exported via NBD.

Container mechanism 'directory':

A copy or snapshot of the container's filesystem state is made
available as a directory. The method is executed inside the user
namespace associated to the container.

2. Restore API

The restore_get_mechanism() method is used to decide on the restore
mechanism. Currently, 'qemu-img' for VMs, and 'directory' or 'tar' for
containers are possible. It is enough to implement one mechanism for
VMs and one mechanism for containers.

Next, methods for extracting the guest and firewall configuration and
the implementations of the restore mechanism via a pair of methods: an
init method, for making the data available to Proxmox VE and a cleanup
method that is called after restore.

2.1. Restore Mechanisms

VM mechanism 'qemu-img':

The backup provider gives a path to the disk image that will be
restored. The path needs to be something 'qemu-img' can deal with,
e.g. can also be an NBD URI or similar.

Container mechanism 'directory':

The backup provider gives the path to a directory with the full
filesystem structure of the container.

Container mechanism 'tar':

The backup provider gives the path to a (potentially compressed) tar
archive with the full filesystem structure of the container.

See the PVE::BackupProvider::Plugin module for the full API
documentation.

======

This series adapts the backup stack in Proxmox VE to allow using the
above API. For QEMU, backup access setup and teardown QMP commands are
implemented to be able to provide access to a consistent disk state to
the backup provider.

The series also provides an example implementation for a backup
provider as a proof-of-concept, exposing the different features.

======

Open questions:

Should the backup provider plugin system also follow the same API
age+version schema with a Custom/ directory for external plugins
derived from the base plugin?

Should the bitmap action be passed directly to the backup provider?
I.e. have 'not-used', 'not-used-removed', 'new', 'used', 'invalid',
instead of only 'none', 'new' and 'reuse'. It makes API slightly more
complicated. Is there any situation where backup provider could care
if bitmap is new, because it was the first or bitmap is new because
previous was invalid? Both cases require the backup provider to do a
full backup.

======

Feedback is very welcome, especially from people wishing to implement
such a backup provider plugin! Please tell me what issues you see with
the proposed API, what would and wouldn't work from your perspective?

======

Dependencies: libpve-storage-perl depends on pve-common. pve-manager,
pve-container and qemu-server all depend on new libpve-storage-perl.
pve-manager also build-depends on the new libpve-storage-perl for its
tests. pve-container depends on new pve-common, i.e. 8.2.9 for the
"tools: run fork: allow running code in parent after fork"
functionality. To keep things clean, pve-manager should also depend on
new pve-container and qemu-server.

======

qemu:

Fiona Ebner (8):
  PVE backup: clean up directly in setup_snapshot_access() when it fails
  PVE backup: factor out helper to clear backup state's bitmap list
  PVE backup: factor out helper to initialize backup state stat struct
  PVE backup: add target ID in backup state
  PVE backup: get device info: allow caller to specify filter for which
    devices use fleecing
  PVE backup: implement backup access setup and teardown API for
    external providers
  PVE backup: implement bitmap support for external backup access
  PVE backup: backup-access api: indicate situation where a bitmap was
    recreated

 pve-backup.c         | 543 +++++++++++++++++++++++++++++++++++++++----
 pve-backup.h         |  16 ++
 qapi/block-core.json |  74 +++++-
 system/runstate.c    |   6 +
 4 files changed, 598 insertions(+), 41 deletions(-)
 create mode 100644 pve-backup.h


common:

Fiona Ebner (1):
  syscall: expose fallocate syscall

 src/PVE/Syscall.pm | 1 +
 1 file changed, 1 insertion(+)


storage:

Fiona Ebner (8):
  add storage_has_feature() helper function
  common: add deallocate helper function
  plugin: introduce new_backup_provider() method
  config api/plugins: let plugins define sensitive properties themselves
  plugin api: bump api version and age
  extract backup config: delegate to backup provider for storages that
    support it
  add backup provider example
  Borg example plugin

 ApiChangeLog                                  |   32 +
 src/PVE/API2/Storage/Config.pm                |    4 +-
 src/PVE/BackupProvider/Makefile               |    3 +
 src/PVE/BackupProvider/Plugin/Base.pm         | 1067 +++++++++++++++++
 src/PVE/BackupProvider/Plugin/Borg.pm         |  460 +++++++
 .../BackupProvider/Plugin/DirectoryExample.pm |  778 ++++++++++++
 src/PVE/BackupProvider/Plugin/Makefile        |    5 +
 src/PVE/Makefile                              |    1 +
 src/PVE/Storage.pm                            |   31 +-
 src/PVE/Storage/BTRFSPlugin.pm                |    1 +
 src/PVE/Storage/CIFSPlugin.pm                 |    1 +
 src/PVE/Storage/CephFSPlugin.pm               |    1 +
 src/PVE/Storage/Common.pm                     |   30 +
 .../Custom/BackupProviderDirExamplePlugin.pm  |  308 +++++
 src/PVE/Storage/Custom/BorgBackupPlugin.pm    |  689 +++++++++++
 src/PVE/Storage/Custom/Makefile               |    6 +
 src/PVE/Storage/DirPlugin.pm                  |    1 +
 src/PVE/Storage/ESXiPlugin.pm                 |    1 +
 src/PVE/Storage/GlusterfsPlugin.pm            |    1 +
 src/PVE/Storage/ISCSIDirectPlugin.pm          |    1 +
 src/PVE/Storage/ISCSIPlugin.pm                |    1 +
 src/PVE/Storage/LVMPlugin.pm                  |    1 +
 src/PVE/Storage/LvmThinPlugin.pm              |    1 +
 src/PVE/Storage/Makefile                      |    1 +
 src/PVE/Storage/NFSPlugin.pm                  |    1 +
 src/PVE/Storage/PBSPlugin.pm                  |    5 +
 src/PVE/Storage/Plugin.pm                     |   37 +
 src/PVE/Storage/RBDPlugin.pm                  |    1 +
 src/PVE/Storage/ZFSPlugin.pm                  |    1 +
 src/PVE/Storage/ZFSPoolPlugin.pm              |    1 +
 30 files changed, 3467 insertions(+), 4 deletions(-)
 create mode 100644 src/PVE/BackupProvider/Makefile
 create mode 100644 src/PVE/BackupProvider/Plugin/Base.pm
 create mode 100644 src/PVE/BackupProvider/Plugin/Borg.pm
 create mode 100644 src/PVE/BackupProvider/Plugin/DirectoryExample.pm
 create mode 100644 src/PVE/BackupProvider/Plugin/Makefile
 create mode 100644 src/PVE/Storage/Custom/BackupProviderDirExamplePlugin.pm
 create mode 100644 src/PVE/Storage/Custom/BorgBackupPlugin.pm
 create mode 100644 src/PVE/Storage/Custom/Makefile


qemu-server:

Fiona Ebner (11):
  backup: keep track of block-node size for fleecing
  backup: fleecing: use exact size when allocating non-raw fleecing
    images
  backup: allow adding fleecing images also for EFI and TPM
  backup: implement backup for external providers
  test: qemu img convert: add test cases for snapshots
  image convert: collect options in hash argument
  image convert: allow caller to specify the format of the source path
  backup: implement restore for external providers
  backup: future-proof checks for QEMU feature support
  backup: support 'missing-recreated' bitmap action
  backup: bitmap action to human: lie about TPM state

 PVE/API2/Qemu.pm                   |  30 +-
 PVE/QemuServer.pm                  | 174 ++++++++++-
 PVE/QemuServer/ImportDisk.pm       |   3 +-
 PVE/VZDump/QemuServer.pm           | 461 ++++++++++++++++++++++++++++-
 test/run_qemu_img_convert_tests.pl |  72 +++--
 5 files changed, 702 insertions(+), 38 deletions(-)


container:

Fiona Ebner (7):
  add LXC::Namespaces module
  backup: implement backup for external providers
  backup: implement restore for external providers
  external restore: don't use 'one-file-system' tar flag when restoring
    from a directory
  create: factor out compression option helper
  restore tar archive: check potentially untrusted archive
  api: add early check against restoring privileged container from
    external source

 src/PVE/API2/LXC.pm       |  14 ++
 src/PVE/LXC/Create.pm     | 273 +++++++++++++++++++++++++++++++++++---
 src/PVE/LXC/Makefile      |   1 +
 src/PVE/LXC/Namespaces.pm |  57 ++++++++
 src/PVE/VZDump/LXC.pm     |  40 +++++-
 5 files changed, 362 insertions(+), 23 deletions(-)
 create mode 100644 src/PVE/LXC/Namespaces.pm


manager:

Fiona Ebner (2):
  ui: backup: also check for backup subtype to classify archive
  backup: implement backup for external providers

 PVE/VZDump.pm                      | 60 ++++++++++++++++++++++++++----
 test/vzdump_new_test.pl            |  3 ++
 www/manager6/Utils.js              | 10 +++--
 www/manager6/grid/BackupView.js    |  4 +-
 www/manager6/storage/BackupView.js |  4 +-
 5 files changed, 66 insertions(+), 15 deletions(-)


Summary over all repositories:
  50 files changed, 5196 insertions(+), 121 deletions(-)

-- 
Generated by git-murpp 0.5.0


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 01/37] PVE backup: clean up directly in setup_snapshot_access() when it fails

[Fiona Ebner]

The only thing that might need to be cleaned up after
setup_snapshot_access() failed is dropping the cbw filter. Do so in
the single branch it matters inside setup_snapshot_access() itself.
This avoids the need that callers of setup_snapshot_access() use
cleanup_snapshot_access() when the call failed.

Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 pve-backup.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/pve-backup.c b/pve-backup.c
index 32352fb5ec..2408f182bc 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -576,6 +576,9 @@ static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
     di->fleecing.snapshot_access =
         bdrv_open(NULL, NULL, snapshot_access_opts, BDRV_O_RDWR | BDRV_O_UNMAP, &local_err);
     if (!di->fleecing.snapshot_access) {
+        bdrv_cbw_drop(di->fleecing.cbw);
+        di->fleecing.cbw = NULL;
+
         error_setg(errp, "setting up snapshot access for fleecing failed: %s",
                    local_err ? error_get_pretty(local_err) : "unknown error");
         return -1;
@@ -629,7 +632,6 @@ static void create_backup_jobs_bh(void *opaque) {
                 error_setg(errp, "%s - setting up snapshot access for fleecing failed: %s",
                            di->device_name,
                            local_err ? error_get_pretty(local_err) : "unknown error");
-                cleanup_snapshot_access(di);
                 bdrv_drained_end(di->bs);
                 break;
             }
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 02/37] PVE backup: factor out helper to clear backup state's bitmap list

[Fiona Ebner]

Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 pve-backup.c | 28 ++++++++++++++++++----------
 1 file changed, 18 insertions(+), 10 deletions(-)

diff --git a/pve-backup.c b/pve-backup.c
index 2408f182bc..915649b5f9 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -811,6 +811,23 @@ err:
     return di_list;
 }
 
+/*
+ * To be called with the backup_state.stat mutex held.
+ */
+static void clear_backup_state_bitmap_list(void) {
+
+    if (backup_state.stat.bitmap_list) {
+        GList *bl = backup_state.stat.bitmap_list;
+        while (bl) {
+            g_free(((PBSBitmapInfo *)bl->data)->drive);
+            g_free(bl->data);
+            bl = g_list_next(bl);
+        }
+        g_list_free(backup_state.stat.bitmap_list);
+        backup_state.stat.bitmap_list = NULL;
+    }
+}
+
 UuidInfo coroutine_fn *qmp_backup(
     const char *backup_file,
     const char *password,
@@ -898,16 +915,7 @@ UuidInfo coroutine_fn *qmp_backup(
     backup_state.stat.reused = 0;
 
     /* clear previous backup's bitmap_list */
-    if (backup_state.stat.bitmap_list) {
-        GList *bl = backup_state.stat.bitmap_list;
-        while (bl) {
-            g_free(((PBSBitmapInfo *)bl->data)->drive);
-            g_free(bl->data);
-            bl = g_list_next(bl);
-        }
-        g_list_free(backup_state.stat.bitmap_list);
-        backup_state.stat.bitmap_list = NULL;
-    }
+    clear_backup_state_bitmap_list();
 
     if (format == BACKUP_FORMAT_PBS) {
         if (!password) {
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 03/37] PVE backup: factor out helper to initialize backup state stat struct

[Fiona Ebner]

Suggested-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 pve-backup.c | 62 ++++++++++++++++++++++++++++++++--------------------
 1 file changed, 38 insertions(+), 24 deletions(-)

diff --git a/pve-backup.c b/pve-backup.c
index 915649b5f9..88a981f81c 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -828,6 +828,43 @@ static void clear_backup_state_bitmap_list(void) {
     }
 }
 
+/*
+ * Initializes most of the backup state 'stat' struct. Note that 'reused' and
+ * 'bitmap_list' are not changed by this function and need to be handled by
+ * the caller. In particular, 'reused' needs to be set before calling this
+ * function.
+ *
+ * To be called with the backup_state.stat mutex held.
+ */
+static void initialize_backup_state_stat(
+    const char *backup_file,
+    uuid_t uuid,
+    size_t total)
+{
+    if (backup_state.stat.error) {
+        error_free(backup_state.stat.error);
+        backup_state.stat.error = NULL;
+    }
+
+    backup_state.stat.start_time = time(NULL);
+    backup_state.stat.end_time = 0;
+
+    if (backup_state.stat.backup_file) {
+        g_free(backup_state.stat.backup_file);
+    }
+    backup_state.stat.backup_file = g_strdup(backup_file);
+
+    uuid_copy(backup_state.stat.uuid, uuid);
+    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
+
+    backup_state.stat.total = total;
+    backup_state.stat.dirty = total - backup_state.stat.reused;
+    backup_state.stat.transferred = 0;
+    backup_state.stat.zero_bytes = 0;
+    backup_state.stat.finishing = false;
+    backup_state.stat.starting = true;
+}
+
 UuidInfo coroutine_fn *qmp_backup(
     const char *backup_file,
     const char *password,
@@ -1070,32 +1107,9 @@ UuidInfo coroutine_fn *qmp_backup(
         }
     }
     /* initialize global backup_state now */
-    /* note: 'reused' and 'bitmap_list' are initialized earlier */
-
-    if (backup_state.stat.error) {
-        error_free(backup_state.stat.error);
-        backup_state.stat.error = NULL;
-    }
-
-    backup_state.stat.start_time = time(NULL);
-    backup_state.stat.end_time = 0;
-
-    if (backup_state.stat.backup_file) {
-        g_free(backup_state.stat.backup_file);
-    }
-    backup_state.stat.backup_file = g_strdup(backup_file);
-
-    uuid_copy(backup_state.stat.uuid, uuid);
-    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
+    initialize_backup_state_stat(backup_file, uuid, total);
     char *uuid_str = g_strdup(backup_state.stat.uuid_str);
 
-    backup_state.stat.total = total;
-    backup_state.stat.dirty = total - backup_state.stat.reused;
-    backup_state.stat.transferred = 0;
-    backup_state.stat.zero_bytes = 0;
-    backup_state.stat.finishing = false;
-    backup_state.stat.starting = true;
-
     qemu_mutex_unlock(&backup_state.stat.lock);
 
     backup_state.speed = (has_speed && speed > 0) ? speed : 0;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 04/37] PVE backup: add target ID in backup state

[Fiona Ebner]

In preparation for allowing multiple backup providers and potentially
multiple targets for a given provider. Each backup target can then
have its own dirty bitmap and there can be additional checks that the
current backup state is actually associated to the expected target.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Add helper function for setting the value.

 pve-backup.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/pve-backup.c b/pve-backup.c
index 88a981f81c..8789a0667a 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -70,6 +70,7 @@ static struct PVEBackupState {
     JobTxn *txn;
     CoMutex backup_mutex;
     CoMutex dump_callback_mutex;
+    char *target_id;
 } backup_state;
 
 static void pvebackup_init(void)
@@ -865,6 +866,16 @@ static void initialize_backup_state_stat(
     backup_state.stat.starting = true;
 }
 
+/*
+ * To be called with the backup_state mutex held.
+ */
+static void backup_state_set_target_id(const char *target_id) {
+    if (backup_state.target_id) {
+        g_free(backup_state.target_id);
+    }
+    backup_state.target_id = g_strdup(target_id);
+}
+
 UuidInfo coroutine_fn *qmp_backup(
     const char *backup_file,
     const char *password,
@@ -904,7 +915,7 @@ UuidInfo coroutine_fn *qmp_backup(
 
     if (backup_state.di_list) {
         error_set(errp, ERROR_CLASS_GENERIC_ERROR,
-                  "previous backup not finished");
+                  "previous backup for target '%s' not finished", backup_state.target_id);
         qemu_co_mutex_unlock(&backup_state.backup_mutex);
         return NULL;
     }
@@ -1122,6 +1133,8 @@ UuidInfo coroutine_fn *qmp_backup(
     backup_state.vmaw = vmaw;
     backup_state.pbs = pbs;
 
+    backup_state_set_target_id("Proxmox");
+
     backup_state.di_list = di_list;
 
     uuid_info = g_malloc0(sizeof(*uuid_info));
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 05/37] PVE backup: get device info: allow caller to specify filter for which devices use fleecing

[Fiona Ebner]

For providing snapshot-access to external backup providers, EFI and
TPM also need an associated fleecing image. The new caller will thus
need a different filter.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 pve-backup.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/pve-backup.c b/pve-backup.c
index 8789a0667a..755f1abcf1 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -719,7 +719,7 @@ static void create_backup_jobs_bh(void *opaque) {
 /*
  * EFI disk and TPM state are small and it's just not worth setting up fleecing for them.
  */
-static bool device_uses_fleecing(const char *device_id)
+static bool fleecing_no_efi_tpm(const char *device_id)
 {
     return strncmp(device_id, "drive-efidisk", 13) && strncmp(device_id, "drive-tpmstate", 14);
 }
@@ -731,7 +731,7 @@ static bool device_uses_fleecing(const char *device_id)
  */
 static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
     const char *devlist,
-    bool fleecing,
+    bool (*device_uses_fleecing)(const char*),
     Error **errp)
 {
     gchar **devs = NULL;
@@ -757,7 +757,7 @@ static GList coroutine_fn GRAPH_RDLOCK *get_device_info(
             di->bs = bs;
             di->device_name = g_strdup(bdrv_get_device_name(bs));
 
-            if (fleecing && device_uses_fleecing(*d)) {
+            if (device_uses_fleecing && device_uses_fleecing(*d)) {
                 g_autofree gchar *fleecing_devid = g_strconcat(*d, "-fleecing", NULL);
                 BlockBackend *fleecing_blk = blk_by_name(fleecing_devid);
                 if (!fleecing_blk) {
@@ -924,7 +924,8 @@ UuidInfo coroutine_fn *qmp_backup(
     format = has_format ? format : BACKUP_FORMAT_VMA;
 
     bdrv_graph_co_rdlock();
-    di_list = get_device_info(devlist, has_fleecing && fleecing, &local_err);
+    di_list = get_device_info(devlist, (has_fleecing && fleecing) ? fleecing_no_efi_tpm : NULL,
+                              &local_err);
     bdrv_graph_co_rdunlock();
     if (local_err) {
         error_propagate(errp, local_err);
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 06/37] PVE backup: implement backup access setup and teardown API for external providers

[Fiona Ebner]

For external backup providers, the state of the VM's disk images at
the time the backup is started is preserved via a snapshot-access
block node. Old data is moved to the fleecing image when new guest
writes come in. The snapshot-access block node, as well as the
associated bitmap in case of incremental backup, will be exported via
NBD to the external provider. The NBD export will be done by the
management layer, the missing functionality is setting up and tearing
down the snapshot-access block nodes, which this patch adds.

It is necessary to also set up fleecing for EFI and TPM disks, so that
old data can be moved out of the way when a new guest write comes in.

There can only be one regular backup or one active backup access at
a time, because both require replacing the original block node of the
drive. Thus the backup state is re-used, and checks are added to
prohibit regular backup while snapshot access is active and vice
versa.

The block nodes added by the backup-access-setup QMP call are not
tracked anywhere else (there is no job they are associated to like for
regular backup). This requires adding a callback for teardown when
QEMU exits, i.e. in qemu_cleanup(). Otherwise, there will be an
assertion failure that the block graph is not empty when QEMU exits
before the backup-access-teardown QMP command is called.

The code for the qmp_backup_access_setup() was based on the existing
qmp_backup() routine.

The return value for the setup QMP command contains information about
the snapshot-access block nodes that can be used by the management
layer to set up the NBD exports.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* setup_snapshot_access() now cleans up after itself, so no need to
  use cleanup_snapshot_access() on the call-side after failure.
* Use clear_backup_state_bitmap_list() helper.
* Use initialize_backup_state_stat() helper.

 pve-backup.c         | 264 ++++++++++++++++++++++++++++++++++++++++++-
 pve-backup.h         |  16 +++
 qapi/block-core.json |  49 ++++++++
 system/runstate.c    |   6 +
 4 files changed, 329 insertions(+), 6 deletions(-)
 create mode 100644 pve-backup.h

diff --git a/pve-backup.c b/pve-backup.c
index 755f1abcf1..091b5bd231 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -1,4 +1,5 @@
 #include "proxmox-backup-client.h"
+#include "pve-backup.h"
 #include "vma.h"
 
 #include "qemu/osdep.h"
@@ -588,6 +589,36 @@ static int setup_snapshot_access(PVEBackupDevInfo *di, Error **errp)
     return 0;
 }
 
+static void setup_all_snapshot_access_bh(void *opaque)
+{
+    assert(!qemu_in_coroutine());
+
+    CoCtxData *data = (CoCtxData*)opaque;
+    Error **errp = (Error**)data->data;
+
+    Error *local_err = NULL;
+
+    GList *l =  backup_state.di_list;
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        bdrv_drained_begin(di->bs);
+
+        if (setup_snapshot_access(di, &local_err) < 0) {
+            bdrv_drained_end(di->bs);
+            error_setg(errp, "%s - setting up snapshot access failed: %s", di->device_name,
+                       local_err ? error_get_pretty(local_err) : "unknown error");
+            break;
+        }
+
+        bdrv_drained_end(di->bs);
+    }
+
+    /* return */
+    aio_co_enter(data->ctx, data->co);
+}
+
 /*
  * backup_job_create can *not* be run from a coroutine, so this can't either.
  * The caller is responsible that backup_mutex is held nonetheless.
@@ -724,6 +755,11 @@ static bool fleecing_no_efi_tpm(const char *device_id)
     return strncmp(device_id, "drive-efidisk", 13) && strncmp(device_id, "drive-tpmstate", 14);
 }
 
+static bool fleecing_all(const char *device_id)
+{
+    return true;
+}
+
 /*
  * Returns a list of device infos, which needs to be freed by the caller. In
  * case of an error, errp will be set, but the returned value might still be a
@@ -839,8 +875,9 @@ static void clear_backup_state_bitmap_list(void) {
  */
 static void initialize_backup_state_stat(
     const char *backup_file,
-    uuid_t uuid,
-    size_t total)
+    uuid_t *uuid,
+    size_t total,
+    bool starting)
 {
     if (backup_state.stat.error) {
         error_free(backup_state.stat.error);
@@ -855,15 +892,19 @@ static void initialize_backup_state_stat(
     }
     backup_state.stat.backup_file = g_strdup(backup_file);
 
-    uuid_copy(backup_state.stat.uuid, uuid);
-    uuid_unparse_lower(uuid, backup_state.stat.uuid_str);
+    if (uuid) {
+        uuid_copy(backup_state.stat.uuid, *uuid);
+        uuid_unparse_lower(*uuid, backup_state.stat.uuid_str);
+    } else {
+        backup_state.stat.uuid_str[0] = '\0';
+    }
 
     backup_state.stat.total = total;
     backup_state.stat.dirty = total - backup_state.stat.reused;
     backup_state.stat.transferred = 0;
     backup_state.stat.zero_bytes = 0;
     backup_state.stat.finishing = false;
-    backup_state.stat.starting = true;
+    backup_state.stat.starting = starting;
 }
 
 /*
@@ -876,6 +917,216 @@ static void backup_state_set_target_id(const char *target_id) {
     backup_state.target_id = g_strdup(target_id);
 }
 
+BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
+    const char *target_id,
+    const char *devlist,
+    Error **errp)
+{
+    assert(qemu_in_coroutine());
+
+    qemu_co_mutex_lock(&backup_state.backup_mutex);
+
+    Error *local_err = NULL;
+    GList *di_list = NULL;
+    GList *l;
+
+    if (backup_state.di_list) {
+        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
+                  "previous backup for target '%s' not finished", backup_state.target_id);
+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
+        return NULL;
+    }
+
+    bdrv_graph_co_rdlock();
+    di_list = get_device_info(devlist, fleecing_all, &local_err);
+    bdrv_graph_co_rdunlock();
+    if (local_err) {
+        error_propagate(errp, local_err);
+        goto err;
+    }
+    assert(di_list);
+
+    size_t total = 0;
+
+    l = di_list;
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        ssize_t size = bdrv_getlength(di->bs);
+        if (size < 0) {
+            error_setg_errno(errp, -size, "bdrv_getlength failed");
+            goto err;
+        }
+        di->size = size;
+        total += size;
+
+        di->completed_ret = INT_MAX;
+    }
+
+    qemu_mutex_lock(&backup_state.stat.lock);
+    backup_state.stat.reused = 0;
+
+    /* clear previous backup's bitmap_list */
+    clear_backup_state_bitmap_list();
+
+    /* starting=false, because there is no associated QEMU job */
+    initialize_backup_state_stat(NULL, NULL, total, false);
+
+    qemu_mutex_unlock(&backup_state.stat.lock);
+
+    backup_state_set_target_id(target_id);
+
+    backup_state.vmaw = NULL;
+    backup_state.pbs = NULL;
+
+    backup_state.di_list = di_list;
+
+    /* Run setup_all_snapshot_access_bh outside of coroutine (in BH) but keep
+    * backup_mutex locked. This is fine, a CoMutex can be held across yield
+    * points, and we'll release it as soon as the BH reschedules us.
+    */
+    CoCtxData waker = {
+        .co = qemu_coroutine_self(),
+        .ctx = qemu_get_current_aio_context(),
+        .data = &local_err,
+    };
+    aio_bh_schedule_oneshot(waker.ctx, setup_all_snapshot_access_bh, &waker);
+    qemu_coroutine_yield();
+
+    if (local_err) {
+        error_propagate(errp, local_err);
+        goto err;
+    }
+
+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
+
+    BackupAccessInfoList *bai_head = NULL, **p_bai_next = &bai_head;
+
+    l = di_list;
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        BackupAccessInfoList *info = g_malloc0(sizeof(*info));
+        info->value = g_malloc0(sizeof(*info->value));
+        info->value->node_name = g_strdup(bdrv_get_node_name(di->fleecing.snapshot_access));
+        info->value->device = g_strdup(di->device_name);
+        info->value->size = di->size;
+
+        *p_bai_next = info;
+        p_bai_next = &info->next;
+    }
+
+    return bai_head;
+
+err:
+
+    l = di_list;
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        g_free(di->device_name);
+        di->device_name = NULL;
+
+        g_free(di);
+    }
+    g_list_free(di_list);
+    backup_state.di_list = NULL;
+
+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
+    return NULL;
+}
+
+/*
+ * Caller needs to hold the backup mutex or the BQL.
+ */
+void backup_access_teardown(void)
+{
+    GList *l = backup_state.di_list;
+
+    qemu_mutex_lock(&backup_state.stat.lock);
+    backup_state.stat.finishing = true;
+    qemu_mutex_unlock(&backup_state.stat.lock);
+
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        if (di->fleecing.snapshot_access) {
+            bdrv_unref(di->fleecing.snapshot_access);
+            di->fleecing.snapshot_access = NULL;
+        }
+        if (di->fleecing.cbw) {
+            bdrv_cbw_drop(di->fleecing.cbw);
+            di->fleecing.cbw = NULL;
+        }
+
+        g_free(di->device_name);
+        di->device_name = NULL;
+
+        g_free(di);
+    }
+    g_list_free(backup_state.di_list);
+    backup_state.di_list = NULL;
+
+    qemu_mutex_lock(&backup_state.stat.lock);
+    backup_state.stat.end_time = time(NULL);
+    backup_state.stat.finishing = false;
+    qemu_mutex_unlock(&backup_state.stat.lock);
+}
+
+// Not done in a coroutine, because bdrv_co_unref() and cbw_drop() would just spawn BHs anyways.
+// Caller needs to hold the backup_state.backup_mutex lock
+static void backup_access_teardown_bh(void *opaque)
+{
+    CoCtxData *data = (CoCtxData*)opaque;
+
+    backup_access_teardown();
+
+    /* return */
+    aio_co_enter(data->ctx, data->co);
+}
+
+void coroutine_fn qmp_backup_access_teardown(const char *target_id, Error **errp)
+{
+    assert(qemu_in_coroutine());
+
+    qemu_co_mutex_lock(&backup_state.backup_mutex);
+
+    if (!backup_state.target_id) { // nothing to do
+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
+        return;
+    }
+
+    /*
+     * Continue with target_id == NULL, used by the callback registered for qemu_cleanup()
+     */
+    if (target_id && strcmp(backup_state.target_id, target_id)) {
+        error_setg(errp, "cannot teardown backup access - got target %s instead of %s",
+                   target_id, backup_state.target_id);
+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
+        return;
+    }
+
+    if (!strcmp(backup_state.target_id, "Proxmox VE")) {
+        error_setg(errp, "cannot teardown backup access for PVE - use backup-cancel instead");
+        qemu_co_mutex_unlock(&backup_state.backup_mutex);
+        return;
+    }
+
+    CoCtxData waker = {
+        .co = qemu_coroutine_self(),
+        .ctx = qemu_get_current_aio_context(),
+    };
+    aio_bh_schedule_oneshot(waker.ctx, backup_access_teardown_bh, &waker);
+    qemu_coroutine_yield();
+
+    qemu_co_mutex_unlock(&backup_state.backup_mutex);
+    return;
+}
+
 UuidInfo coroutine_fn *qmp_backup(
     const char *backup_file,
     const char *password,
@@ -1119,7 +1370,7 @@ UuidInfo coroutine_fn *qmp_backup(
         }
     }
     /* initialize global backup_state now */
-    initialize_backup_state_stat(backup_file, uuid, total);
+    initialize_backup_state_stat(backup_file, &uuid, total, true);
     char *uuid_str = g_strdup(backup_state.stat.uuid_str);
 
     qemu_mutex_unlock(&backup_state.stat.lock);
@@ -1298,5 +1549,6 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
     ret->pbs_masterkey = true;
     ret->backup_max_workers = true;
     ret->backup_fleecing = true;
+    ret->backup_access_api = true;
     return ret;
 }
diff --git a/pve-backup.h b/pve-backup.h
new file mode 100644
index 0000000000..4033bc848f
--- /dev/null
+++ b/pve-backup.h
@@ -0,0 +1,16 @@
+/*
+ * Bacup code used by Proxmox VE
+ *
+ * Copyright (C) Proxmox Server Solutions
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#ifndef PVE_BACKUP_H
+#define PVE_BACKUP_H
+
+void backup_access_teardown(void);
+
+#endif /* PVE_BACKUP_H */
diff --git a/qapi/block-core.json b/qapi/block-core.json
index c581f1f238..3f092221ce 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1019,6 +1019,9 @@
 #
 # @pbs-library-version: Running version of libproxmox-backup-qemu0 library.
 #
+# @backup-access-api: Whether backup access API for external providers is
+#     supported or not.
+#
 # @backup-fleecing: Whether backup fleecing is supported or not.
 #
 # @backup-max-workers: Whether the 'max-workers' @BackupPerf setting is
@@ -1032,6 +1035,7 @@
             'pbs-dirty-bitmap-migration': 'bool',
             'pbs-masterkey': 'bool',
             'pbs-library-version': 'str',
+            'backup-access-api': 'bool',
             'backup-fleecing': 'bool',
             'backup-max-workers': 'bool' } }
 
@@ -1098,6 +1102,51 @@
 ##
 { 'command': 'query-pbs-bitmap-info', 'returns': ['PBSBitmapInfo'] }
 
+##
+# @BackupAccessInfo:
+#
+# Info associated to a snapshot access for backup.  For more information about
+# the bitmap see @BackupAccessBitmapMode.
+#
+# @node-name: the block node name of the snapshot-access node.
+#
+# @device: the device on top of which the snapshot access was created.
+#
+# @size: the size of the block device in bytes.
+#
+##
+{ 'struct': 'BackupAccessInfo',
+  'data': { 'node-name': 'str', 'device': 'str', 'size': 'size' } }
+
+##
+# @backup-access-setup:
+#
+# Set up snapshot access to VM drives for an external backup provider.  No other
+# backup or backup access can be done before tearing down the backup access.
+#
+# @target-id: the unique ID of the backup target.
+#
+# @devlist: list of block device names (separated by ',', ';' or ':'). By
+#     default the backup includes all writable block devices.
+#
+# Returns: a list of @BackupAccessInfo, one for each device.
+#
+##
+{ 'command': 'backup-access-setup',
+  'data': { 'target-id': 'str', '*devlist': 'str' },
+  'returns': [ 'BackupAccessInfo' ], 'coroutine': true }
+
+##
+# @backup-access-teardown:
+#
+# Tear down previously setup snapshot access for the same target.
+#
+# @target-id: the ID of the backup target.
+#
+##
+{ 'command': 'backup-access-teardown', 'data': { 'target-id': 'str' },
+  'coroutine': true }
+
 ##
 # @BlockDeviceTimedStats:
 #
diff --git a/system/runstate.c b/system/runstate.c
index c2c9afa905..6f93d7c2fb 100644
--- a/system/runstate.c
+++ b/system/runstate.c
@@ -60,6 +60,7 @@
 #include "sysemu/sysemu.h"
 #include "sysemu/tpm.h"
 #include "trace.h"
+#include "pve-backup.h"
 
 static NotifierList exit_notifiers =
     NOTIFIER_LIST_INITIALIZER(exit_notifiers);
@@ -920,6 +921,11 @@ void qemu_cleanup(int status)
      * requests happening from here on anyway.
      */
     bdrv_drain_all_begin();
+    /*
+     * The backup access is set up by a QMP command, but is neither owned by a monitor nor
+     * associated to a BlockBackend. Need to tear it down manually here.
+     */
+    backup_access_teardown();
     job_cancel_sync_all();
     bdrv_close_all();
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 07/37] PVE backup: implement bitmap support for external backup access

[Fiona Ebner]

There can be one dirty bitmap for each backup target ID (which are
tracked in the backup_access_bitmaps hash table). The QMP user can
specify the ID of the bitmap it likes to use. This ID is then compared
to the current one for the given target. If they match, the bitmap is
re-used (should it still exist on the drive, otherwise re-created). If
there is a mismatch, the old bitmap is removed and a new one is
created.

The return value of the QMP command includes information about what
bitmap action was taken. Similar to what the query-backup QMP command
returns for regular backup. It also includes the bitmap name and
associated block node, so the management layer can then set up an NBD
export with the bitmap.

While the backup access is active, a background bitmap is also
required. This is necessary to implement bitmap handling according to
the original reference [0]. In particular:

- in the error case, new writes since the backup access was set up are
  in the background bitmap. Because of failure, the previously tracked
  writes from the backup access bitmap are still required too. Thus,
  the bitmap is merged with the background bitmap to get all new
  writes since the last backup.

- in the success case, continue tracking for the next incremental
  backup in the backup access bitmap. New writes since the backup
  access was set up are in the background bitmap. Because the backup
  was successfully, clear the backup access bitmap and merge back the
  background bitmap to get only the new writes.

Since QEMU cannot know if the backup was successful or not (except if
failure already happens during the setup QMP command), the management
layer needs to tell it via the teardown QMP command.

The bitmap action is also recorded in the device info now.

[0]: https://lore.kernel.org/qemu-devel/b68833dd-8864-4d72-7c61-c134a9835036@ya.ru/

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 pve-backup.c         | 175 ++++++++++++++++++++++++++++++++++++++++++-
 pve-backup.h         |   2 +-
 qapi/block-core.json |  22 +++++-
 system/runstate.c    |   2 +-
 4 files changed, 193 insertions(+), 8 deletions(-)

diff --git a/pve-backup.c b/pve-backup.c
index 091b5bd231..366c4c198d 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -15,6 +15,7 @@
 #include "qapi/qmp/qdict.h"
 #include "qapi/qmp/qerror.h"
 #include "qemu/cutils.h"
+#include "qemu/error-report.h"
 
 #if defined(CONFIG_MALLOC_TRIM)
 #include <malloc.h>
@@ -41,6 +42,7 @@
  */
 
 const char *PBS_BITMAP_NAME = "pbs-incremental-dirty-bitmap";
+const char *BACKGROUND_BITMAP_NAME = "backup-access-background-bitmap";
 
 static struct PVEBackupState {
     struct {
@@ -72,6 +74,7 @@ static struct PVEBackupState {
     CoMutex backup_mutex;
     CoMutex dump_callback_mutex;
     char *target_id;
+    GHashTable *backup_access_bitmaps; // key=target_id, value=bitmap_name
 } backup_state;
 
 static void pvebackup_init(void)
@@ -99,6 +102,8 @@ typedef struct PVEBackupDevInfo {
     char* device_name;
     int completed_ret; // INT_MAX if not completed
     BdrvDirtyBitmap *bitmap;
+    BdrvDirtyBitmap *background_bitmap; // used for external backup access
+    PBSBitmapAction bitmap_action;
     BlockDriverState *target;
     BlockJob *job;
 } PVEBackupDevInfo;
@@ -362,6 +367,67 @@ static void coroutine_fn pvebackup_co_complete_stream(void *opaque)
     qemu_co_mutex_unlock(&backup_state.backup_mutex);
 }
 
+/*
+ * New writes since the backup access was set up are in the background bitmap. Because of failure,
+ * the previously tracked writes in di->bitmap are still required too. Thus, merge with the
+ * background bitmap to get all new writes since the last backup.
+ */
+static void handle_backup_access_bitmaps_in_error_case(PVEBackupDevInfo *di)
+{
+    Error *local_err = NULL;
+
+    if (di->bs && di->background_bitmap) {
+        bdrv_drained_begin(di->bs);
+        if (di->bitmap) {
+            bdrv_enable_dirty_bitmap(di->bitmap);
+            if (!bdrv_merge_dirty_bitmap(di->bitmap, di->background_bitmap, NULL, &local_err)) {
+                warn_report("backup access: %s - could not merge bitmaps in error path - %s",
+                            di->device_name,
+                            local_err ? error_get_pretty(local_err) : "unknown error");
+                /*
+                 * Could not merge, drop original bitmap too.
+                 */
+                bdrv_release_dirty_bitmap(di->bitmap);
+            }
+        } else {
+            warn_report("backup access: %s - expected bitmap not present", di->device_name);
+        }
+        bdrv_release_dirty_bitmap(di->background_bitmap);
+        bdrv_drained_end(di->bs);
+    }
+}
+
+/*
+ * Continue tracking for next incremental backup in di->bitmap. New writes since the backup access
+ * was set up are in the background bitmap. Because the backup was successful, clear di->bitmap and
+ * merge back the background bitmap to get only the new writes.
+ */
+static void handle_backup_access_bitmaps_after_success(PVEBackupDevInfo *di)
+{
+    Error *local_err = NULL;
+
+    if (di->bs && di->background_bitmap) {
+        bdrv_drained_begin(di->bs);
+        if (di->bitmap) {
+            bdrv_enable_dirty_bitmap(di->bitmap);
+            bdrv_clear_dirty_bitmap(di->bitmap, NULL);
+            if (!bdrv_merge_dirty_bitmap(di->bitmap, di->background_bitmap, NULL, &local_err)) {
+                warn_report("backup access: %s - could not merge bitmaps after backup - %s",
+                            di->device_name,
+                            local_err ? error_get_pretty(local_err) : "unknown error");
+                /*
+                 * Could not merge, drop original bitmap too.
+                 */
+                bdrv_release_dirty_bitmap(di->bitmap);
+            }
+        } else {
+            warn_report("backup access: %s - expected bitmap not present", di->device_name);
+        }
+        bdrv_release_dirty_bitmap(di->background_bitmap);
+        bdrv_drained_end(di->bs);
+    }
+}
+
 static void cleanup_snapshot_access(PVEBackupDevInfo *di)
 {
     if (di->fleecing.snapshot_access) {
@@ -605,6 +671,21 @@ static void setup_all_snapshot_access_bh(void *opaque)
 
         bdrv_drained_begin(di->bs);
 
+        if (di->bitmap) {
+            BdrvDirtyBitmap *background_bitmap =
+                bdrv_create_dirty_bitmap(di->bs, PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE,
+                                         BACKGROUND_BITMAP_NAME, &local_err);
+            if (!background_bitmap) {
+                error_setg(errp, "%s - creating background bitmap for backup access failed: %s",
+                           di->device_name,
+                           local_err ? error_get_pretty(local_err) : "unknown error");
+                bdrv_drained_end(di->bs);
+                break;
+            }
+            di->background_bitmap = background_bitmap;
+            bdrv_disable_dirty_bitmap(di->bitmap);
+        }
+
         if (setup_snapshot_access(di, &local_err) < 0) {
             bdrv_drained_end(di->bs);
             error_setg(errp, "%s - setting up snapshot access failed: %s", di->device_name,
@@ -920,6 +1001,7 @@ static void backup_state_set_target_id(const char *target_id) {
 BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
     const char *target_id,
     const char *devlist,
+    const char *bitmap_name,
     Error **errp)
 {
     assert(qemu_in_coroutine());
@@ -970,6 +1052,77 @@ BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
     /* clear previous backup's bitmap_list */
     clear_backup_state_bitmap_list();
 
+    if (!backup_state.backup_access_bitmaps) {
+        backup_state.backup_access_bitmaps =
+            g_hash_table_new_full(g_str_hash, g_str_equal, free, free);
+    }
+
+    /* create bitmaps if requested */
+    l = di_list;
+    while (l) {
+        PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
+        l = g_list_next(l);
+
+        di->block_size = PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE;
+
+        PBSBitmapAction action = PBS_BITMAP_ACTION_NOT_USED;
+        size_t dirty = di->size;
+
+        const char *old_bitmap_name =
+            (const char*)g_hash_table_lookup(backup_state.backup_access_bitmaps, target_id);
+
+        bool same_bitmap_name =
+            old_bitmap_name && bitmap_name && strcmp(bitmap_name, old_bitmap_name) == 0;
+
+        if (old_bitmap_name && !same_bitmap_name) {
+            BdrvDirtyBitmap *old_bitmap = bdrv_find_dirty_bitmap(di->bs, old_bitmap_name);
+            if (!old_bitmap) {
+                warn_report("setup backup access: expected old bitmap '%s' not found for drive "
+                            "'%s'", old_bitmap_name, di->device_name);
+            } else {
+                g_hash_table_remove(backup_state.backup_access_bitmaps, target_id);
+                bdrv_release_dirty_bitmap(old_bitmap);
+                action = PBS_BITMAP_ACTION_NOT_USED_REMOVED;
+            }
+        }
+
+        BdrvDirtyBitmap *bitmap = NULL;
+        if (bitmap_name) {
+            bitmap = bdrv_find_dirty_bitmap(di->bs, bitmap_name);
+            if (!bitmap) {
+                bitmap = bdrv_create_dirty_bitmap(di->bs, PROXMOX_BACKUP_DEFAULT_CHUNK_SIZE,
+                                                  bitmap_name, errp);
+                if (!bitmap) {
+                    qemu_mutex_unlock(&backup_state.stat.lock);
+                    goto err;
+                }
+                bdrv_set_dirty_bitmap(bitmap, 0, di->size);
+                action = PBS_BITMAP_ACTION_NEW;
+            } else {
+                /* track clean chunks as reused */
+                dirty = MIN(bdrv_get_dirty_count(bitmap), di->size);
+                backup_state.stat.reused += di->size - dirty;
+                action = PBS_BITMAP_ACTION_USED;
+            }
+
+            if (!same_bitmap_name) {
+                g_hash_table_insert(backup_state.backup_access_bitmaps,
+                                    strdup(target_id), strdup(bitmap_name));
+            }
+
+        }
+
+        PBSBitmapInfo *info = g_malloc(sizeof(*info));
+        info->drive = g_strdup(di->device_name);
+        info->action = action;
+        info->size = di->size;
+        info->dirty = dirty;
+        backup_state.stat.bitmap_list = g_list_append(backup_state.stat.bitmap_list, info);
+
+        di->bitmap = bitmap;
+        di->bitmap_action = action;
+    }
+
     /* starting=false, because there is no associated QEMU job */
     initialize_backup_state_stat(NULL, NULL, total, false);
 
@@ -1013,6 +1166,12 @@ BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
         info->value->node_name = g_strdup(bdrv_get_node_name(di->fleecing.snapshot_access));
         info->value->device = g_strdup(di->device_name);
         info->value->size = di->size;
+        if (bitmap_name) {
+            info->value->bitmap_node_name = g_strdup(bdrv_get_node_name(di->bs));
+            info->value->bitmap_name = g_strdup(bitmap_name);
+            info->value->bitmap_action = di->bitmap_action;
+            info->value->has_bitmap_action = true;
+        }
 
         *p_bai_next = info;
         p_bai_next = &info->next;
@@ -1027,6 +1186,8 @@ err:
         PVEBackupDevInfo *di = (PVEBackupDevInfo *)l->data;
         l = g_list_next(l);
 
+        handle_backup_access_bitmaps_in_error_case(di);
+
         g_free(di->device_name);
         di->device_name = NULL;
 
@@ -1042,7 +1203,7 @@ err:
 /*
  * Caller needs to hold the backup mutex or the BQL.
  */
-void backup_access_teardown(void)
+void backup_access_teardown(bool success)
 {
     GList *l = backup_state.di_list;
 
@@ -1063,6 +1224,12 @@ void backup_access_teardown(void)
             di->fleecing.cbw = NULL;
         }
 
+        if (success) {
+            handle_backup_access_bitmaps_after_success(di);
+        } else {
+            handle_backup_access_bitmaps_in_error_case(di);
+        }
+
         g_free(di->device_name);
         di->device_name = NULL;
 
@@ -1083,13 +1250,13 @@ static void backup_access_teardown_bh(void *opaque)
 {
     CoCtxData *data = (CoCtxData*)opaque;
 
-    backup_access_teardown();
+    backup_access_teardown(*((bool*)data->data));
 
     /* return */
     aio_co_enter(data->ctx, data->co);
 }
 
-void coroutine_fn qmp_backup_access_teardown(const char *target_id, Error **errp)
+void coroutine_fn qmp_backup_access_teardown(const char *target_id, bool success, Error **errp)
 {
     assert(qemu_in_coroutine());
 
@@ -1119,6 +1286,7 @@ void coroutine_fn qmp_backup_access_teardown(const char *target_id, Error **errp
     CoCtxData waker = {
         .co = qemu_coroutine_self(),
         .ctx = qemu_get_current_aio_context(),
+        .data = &success,
     };
     aio_bh_schedule_oneshot(waker.ctx, backup_access_teardown_bh, &waker);
     qemu_coroutine_yield();
@@ -1319,6 +1487,7 @@ UuidInfo coroutine_fn *qmp_backup(
             }
 
             di->dev_id = dev_id;
+            di->bitmap_action = action;
 
             PBSBitmapInfo *info = g_malloc(sizeof(*info));
             info->drive = g_strdup(di->device_name);
diff --git a/pve-backup.h b/pve-backup.h
index 4033bc848f..9ebeef7c8f 100644
--- a/pve-backup.h
+++ b/pve-backup.h
@@ -11,6 +11,6 @@
 #ifndef PVE_BACKUP_H
 #define PVE_BACKUP_H
 
-void backup_access_teardown(void);
+void backup_access_teardown(bool success);
 
 #endif /* PVE_BACKUP_H */
diff --git a/qapi/block-core.json b/qapi/block-core.json
index 3f092221ce..63a5be0f3f 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1114,9 +1114,17 @@
 #
 # @size: the size of the block device in bytes.
 #
+# @bitmap-node-name: the block node name the dirty bitmap is associated to.
+#
+# @bitmap-name: the name of the dirty bitmap associated to the backup access.
+#
+# @bitmap-action: the action taken on the dirty bitmap.
+#
 ##
 { 'struct': 'BackupAccessInfo',
-  'data': { 'node-name': 'str', 'device': 'str', 'size': 'size' } }
+  'data': { 'node-name': 'str', 'device': 'str', 'size': 'size',
+            '*bitmap-node-name': 'str', '*bitmap-name': 'str',
+            '*bitmap-action': 'PBSBitmapAction' } }
 
 ##
 # @backup-access-setup:
@@ -1129,11 +1137,16 @@
 # @devlist: list of block device names (separated by ',', ';' or ':'). By
 #     default the backup includes all writable block devices.
 #
+# @bitmap-name: use/create a bitmap with this name. Re-using the same name
+#     allows for making incremental backups. Check the @bitmap-action in the
+#     result to see if you can actually re-use the bitmap or if it had to be
+#     newly created.
+#
 # Returns: a list of @BackupAccessInfo, one for each device.
 #
 ##
 { 'command': 'backup-access-setup',
-  'data': { 'target-id': 'str', '*devlist': 'str' },
+  'data': { 'target-id': 'str', '*devlist': 'str', '*bitmap-name': 'str' },
   'returns': [ 'BackupAccessInfo' ], 'coroutine': true }
 
 ##
@@ -1143,8 +1156,11 @@
 #
 # @target-id: the ID of the backup target.
 #
+# @success: whether the backup done by the external provider was successful.
+#
 ##
-{ 'command': 'backup-access-teardown', 'data': { 'target-id': 'str' },
+{ 'command': 'backup-access-teardown',
+  'data': { 'target-id': 'str', 'success': 'bool' },
   'coroutine': true }
 
 ##
diff --git a/system/runstate.c b/system/runstate.c
index 6f93d7c2fb..ef3277930f 100644
--- a/system/runstate.c
+++ b/system/runstate.c
@@ -925,7 +925,7 @@ void qemu_cleanup(int status)
      * The backup access is set up by a QMP command, but is neither owned by a monitor nor
      * associated to a BlockBackend. Need to tear it down manually here.
      */
-    backup_access_teardown();
+    backup_access_teardown(false);
     job_cancel_sync_all();
     bdrv_close_all();
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu v6 08/37] PVE backup: backup-access api: indicate situation where a bitmap was recreated

[Fiona Ebner]

The backup-access api keeps track of what bitmap names got used for
which devices and thus knows when a bitmap went missing. Propagate
this information to the QMP user with a new 'missing-recreated'
variant for the taken bitmap action.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 pve-backup.c         | 6 +++++-
 qapi/block-core.json | 9 ++++++++-
 2 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/pve-backup.c b/pve-backup.c
index 366c4c198d..26fe76b76d 100644
--- a/pve-backup.c
+++ b/pve-backup.c
@@ -1097,7 +1097,11 @@ BackupAccessInfoList *coroutine_fn qmp_backup_access_setup(
                     goto err;
                 }
                 bdrv_set_dirty_bitmap(bitmap, 0, di->size);
-                action = PBS_BITMAP_ACTION_NEW;
+                if (same_bitmap_name) {
+                    action = PBS_BITMAP_ACTION_MISSING_RECREATED;
+                } else {
+                    action = PBS_BITMAP_ACTION_NEW;
+                }
             } else {
                 /* track clean chunks as reused */
                 dirty = MIN(bdrv_get_dirty_count(bitmap), di->size);
diff --git a/qapi/block-core.json b/qapi/block-core.json
index 63a5be0f3f..fa6166ab79 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1067,9 +1067,16 @@
 #           base snapshot did not match the base given for the current job or
 #           the crypt mode has changed.
 #
+# @missing-recreated: A bitmap for incremental backup was expected to be
+#     present, but was missing and thus got recreated. For example, this can
+#     happen if the drive was re-attached or if the bitmap was deleted for some
+#     other reason. PBS does not currently keep track of this; the backup-access
+#     mechanism does.
+#
 ##
 { 'enum': 'PBSBitmapAction',
-  'data': ['not-used', 'not-used-removed', 'new', 'used', 'invalid'] }
+  'data': ['not-used', 'not-used-removed', 'new', 'used', 'invalid',
+           'missing-recreated'] }
 
 ##
 # @PBSBitmapInfo:
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH common v6 09/37] syscall: expose fallocate syscall

[Fiona Ebner]

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/Syscall.pm | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/PVE/Syscall.pm b/src/PVE/Syscall.pm
index 9ef3d5d..f3193a3 100644
--- a/src/PVE/Syscall.pm
+++ b/src/PVE/Syscall.pm
@@ -19,6 +19,7 @@ BEGIN {
 	mknod => &SYS_mknod,
 	faccessat => &SYS_faccessat,
 	setresuid => &SYS_setresuid,
+	fallocate => &SYS_fallocate,
 	fchownat => &SYS_fchownat,
 	mount => &SYS_mount,
 	renameat2 => &SYS_renameat2,
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 10/37] add storage_has_feature() helper function

[Fiona Ebner]

Which looks up whether a storage supports a given feature in its
'plugindata'. This is intentionally kept simple and not implemented
as a plugin method for now. Should it ever become more complex
requiring plugins to override the default implementation, it can
later be changed to a method.

Suggested-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/Storage.pm        |  8 ++++++++
 src/PVE/Storage/Plugin.pm | 10 ++++++++++
 2 files changed, 18 insertions(+)

diff --git a/src/PVE/Storage.pm b/src/PVE/Storage.pm
index c5d4ff8..8cbfb4f 100755
--- a/src/PVE/Storage.pm
+++ b/src/PVE/Storage.pm
@@ -213,6 +213,14 @@ sub storage_check_enabled {
     return storage_check_node($cfg, $storeid, $node, $noerr);
 }
 
+sub storage_has_feature {
+    my ($cfg, $storeid, $feature) = @_;
+
+    my $scfg = storage_config($cfg, $storeid);
+
+    return PVE::Storage::Plugin::storage_has_feature($scfg->{type}, $feature);
+}
+
 # storage_can_replicate:
 # return true if storage supports replication
 # (volumes allocated with vdisk_alloc() has replication feature)
diff --git a/src/PVE/Storage/Plugin.pm b/src/PVE/Storage/Plugin.pm
index 65cf43f..80daeea 100644
--- a/src/PVE/Storage/Plugin.pm
+++ b/src/PVE/Storage/Plugin.pm
@@ -246,6 +246,16 @@ sub dirs_hash_to_string {
     return join(',', map { "$_=$hash->{$_}" } sort keys %$hash);
 }
 
+sub storage_has_feature {
+    my ($type, $feature) = @_;
+
+    my $data = $defaultData->{plugindata}->{$type};
+    if (my $features = $data->{features}) {
+	return $features->{$feature};
+    }
+    return;
+}
+
 sub default_format {
     my ($scfg) = @_;
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 11/37] common: add deallocate helper function

[Fiona Ebner]

For punching holes via fallocate. This will be useful for the external
backup provider API to discard parts of the source. The 'file-handle'
mechanism there uses a fuse mount, which does not implement the
BLKDISCARD ioctl, but does implement fallocate.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/Storage/Common.pm | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

diff --git a/src/PVE/Storage/Common.pm b/src/PVE/Storage/Common.pm
index 3ae20dd..32ed26a 100644
--- a/src/PVE/Storage/Common.pm
+++ b/src/PVE/Storage/Common.pm
@@ -3,6 +3,13 @@ package PVE::Storage::Common;
 use strict;
 use warnings;
 
+use PVE::Syscall;
+
+use constant {
+    FALLOC_FL_KEEP_SIZE => 0x01, # see linux/falloc.h
+    FALLOC_FL_PUNCH_HOLE => 0x02, # see linux/falloc.h
+};
+
 =pod
 
 =head1 NAME
@@ -51,4 +58,27 @@ sub align_size_up : prototype($$) {
     return $aligned_size;
 }
 
+=pod
+
+=head3 deallocate
+
+    deallocate($file_handle, $offset, $length)
+
+Deallocates the range with C<$length> many bytes starting from offset C<$offset>
+for the file associated to the file handle C<$file_handle>. Dies on failure.
+
+=cut
+
+sub deallocate : prototype($$$) {
+    my ($file_handle, $offset, $length) = @_;
+
+    my $mode = FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE;
+    $offset = int($offset);
+    $length = int($length);
+
+    if (syscall(PVE::Syscall::fallocate, fileno($file_handle), $mode, $offset, $length) != 0) {
+	die "fallocate: punch hole failed (offset: $offset, length: $length) - $!\n";
+    }
+}
+
 1;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 12/37] plugin: introduce new_backup_provider() method

[Fiona Ebner]

The new_backup_provider() method can be used by storage plugins for
external backup providers. If the method returns a provider, Proxmox
VE will use callbacks to that provider for backups and restore instead
of using its usual backup/restore mechanisms.

The backup provider API is split into two parts, both of which again
need different implementations for VM and LXC guests:

1. Backup API

In Proxmox VE, a backup job consists of backup tasks for individual
guests. There are methods for initialization and cleanup of the job,
i.e. job_init() and job_cleanup() and for each guest backup, i.e.
backup_init() and backup_cleanup().

The backup_get_mechanism() method is used to decide on the backup
mechanism. Currently, 'file-handle' or 'nbd' for VMs, and 'directory'
for containers is possible. The method also let's the plugin indicate
whether to use a bitmap for incremental VM backup or not. It is enough
to implement one mechanism for VMs and one mechanism for containers.

Next, there are methods for backing up the guest's configuration and
data, backup_vm() for VM backup and backup_container() for container
backup, with the latter running

Finally, some helpers like getting the provider name or volume ID for
the backup target, as well as for handling the backup log.

The backup transaction looks as follows:

First, job_init() is called that can be used to check backup server
availability and prepare the connection. Then for each guest
backup_init() followed by backup_vm() or backup_container() and finally
backup_cleanup(). Afterwards job_cleanup() is called. For containers,
there is an additional backup_container_prepare() call while still
privileged. The actual backup_container() call happens as the
(unprivileged) container root user, so that the file owner and group IDs
match the container's perspective.

1.1 Backup Mechanisms

VM:

Access to the data on the VM's disk from the time the backup started
is made available via a so-called "snapshot access". This is either
the full image, or in case a bitmap is used, the dirty parts of the
image since the last time the bitmap was used for a successful backup.
Reading outside of the dirty parts will result in an error. After
backing up each part of the disk, it should be discarded in the export
to avoid unnecessary space usage on the Proxmox VE side (there is an
associated fleecing image).

VM mechanism 'file-handle':

The snapshot access is exposed via a file descriptor. A subroutine to
read the dirty regions for incremental backup is provided as well.

VM mechanism 'nbd':

The snapshot access and, if used, bitmap are exported via NBD.

Container mechanism 'directory':

A copy or snapshot of the container's filesystem state is made
available as a directory. The method is executed inside the user
namespace associated to the container.

2. Restore API

The restore_get_mechanism() method is used to decide on the restore
mechanism. Currently, 'qemu-img' for VMs, and 'directory' or 'tar' for
containers are possible. It is enough to implement one mechanism for
VMs and one mechanism for containers.

Next, methods for extracting the guest and firewall configuration and
the implementations of the restore mechanism via a pair of methods: an
init method, for making the data available to Proxmox VE and a cleanup
method that is called after restore.

2.1. Restore Mechanisms

VM mechanism 'qemu-img':

The backup provider gives a path to the disk image that will be
restored. The path needs to be something 'qemu-img' can deal with,
e.g. can also be an NBD URI or similar.

Container mechanism 'directory':

The backup provider gives the path to a directory with the full
filesystem structure of the container.

Container mechanism 'tar':

The backup provider gives the path to a (potentially compressed) tar
archive with the full filesystem structure of the container.

See the PVE::BackupProvider::Plugin module for the full API
documentation.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Remove superfluous $storeid parameter from methods. It's already
  part of $self.
* Document that volume for VM image restore will be treated as being
  in raw format.
* Document that methods for extracting configurations are called
  without restore_*_init() first.
* Rename restore_get_*_config() methods to archive_get_*_config().
* Replace job/backup hooks with dedicated methods.
* Remove backup_get_archive_name() method. The archive name is now
  returned as part of the backup_init() method.
* Supersede backup_get_task_size() method and return
  extensible-in-the-future stats as part of the backup_cleanup()
  method.

 src/PVE/BackupProvider/Makefile        |    3 +
 src/PVE/BackupProvider/Plugin/Base.pm  | 1067 ++++++++++++++++++++++++
 src/PVE/BackupProvider/Plugin/Makefile |    5 +
 src/PVE/Makefile                       |    1 +
 src/PVE/Storage.pm                     |    8 +
 src/PVE/Storage/Plugin.pm              |   15 +
 6 files changed, 1099 insertions(+)
 create mode 100644 src/PVE/BackupProvider/Makefile
 create mode 100644 src/PVE/BackupProvider/Plugin/Base.pm
 create mode 100644 src/PVE/BackupProvider/Plugin/Makefile

diff --git a/src/PVE/BackupProvider/Makefile b/src/PVE/BackupProvider/Makefile
new file mode 100644
index 0000000..f018cef
--- /dev/null
+++ b/src/PVE/BackupProvider/Makefile
@@ -0,0 +1,3 @@
+.PHONY: install
+install:
+	make -C Plugin install
diff --git a/src/PVE/BackupProvider/Plugin/Base.pm b/src/PVE/BackupProvider/Plugin/Base.pm
new file mode 100644
index 0000000..424feb7
--- /dev/null
+++ b/src/PVE/BackupProvider/Plugin/Base.pm
@@ -0,0 +1,1067 @@
+package PVE::BackupProvider::Plugin::Base;
+
+use strict;
+use warnings;
+
+=pod
+
+=head1 NAME
+
+PVE::BackupProvider::Plugin::Base - Base Plugin for Backup Provider API
+
+=head1 SYNOPSIS
+
+    use base qw(PVE::BackupProvider::Plugin::Base);
+
+=head1 DESCRIPTION
+
+This module serves as the base for any module implementing the API that Proxmox
+VE uses to interface with external backup providers. The API is used for
+creating and restoring backups. A backup provider also needs to provide a
+storage plugin for integration with the front-end. The API here is used by the
+backup stack in the backend.
+
+1. Backup API
+
+In Proxmox VE, a backup job consists of backup tasks for individual guests.
+There are methods for initialization and cleanup of the job, i.e. job_init() and
+job_cleanup() and for each guest backup, i.e. backup_init() and
+backup_cleanup().
+
+The backup_get_mechanism() method is used to decide on the backup mechanism.
+Currently, 'file-handle' or 'nbd' for VMs, and 'directory' for containers is
+possible. The method also let's the plugin indicate whether to use a bitmap for
+incremental VM backup or not. It is enough to implement one mechanism for VMs
+and one mechanism for containers.
+
+Next, there are methods for backing up the guest's configuration and data,
+backup_vm() for VM backup and backup_container() for container backup.
+
+Finally, some helpers like provider_name() for getting the name of the backup
+provider and backup_handle_log_file() for handling the backup task log.
+
+The backup transaction looks as follows:
+
+First, job_init() is called that can be used to check backup server availability
+and prepare the connection. Then for each guest backup_init() followed by
+backup_vm() or backup_container() and finally backup_cleanup(). Afterwards
+job_cleanup() is called. For containers, there is an additional
+backup_container_prepare() call while still privileged. The actual
+backup_container() call happens as the (unprivileged) container root user, so
+that the file owner and group IDs match the container's perspective.
+
+1.1 Backup Mechanisms
+
+VM:
+
+Access to the data on the VM's disk from the time the backup started is made
+available via a so-called "snapshot access". This is either the full image, or
+in case a bitmap is used, the dirty parts of the image since the last time the
+bitmap was used for a successful backup. Reading outside of the dirty parts will
+result in an error. After backing up each part of the disk, it should be
+discarded in the export to avoid unnecessary space usage on the Proxmox VE side
+(there is an associated fleecing image).
+
+VM mechanism 'file-handle':
+
+The snapshot access is exposed via a file descriptor. A subroutine to read the
+dirty regions for incremental backup is provided as well.
+
+VM mechanism 'nbd':
+
+The snapshot access and, if used, bitmap are exported via NBD.
+
+Container mechanism 'directory':
+
+A copy or snapshot of the container's filesystem state is made available as a
+directory.
+
+2. Restore API
+
+The restore_get_mechanism() method is used to decide on the restore mechanism.
+Currently, 'qemu-img' for VMs, and 'directory' or 'tar' for containers are
+possible. It is enough to implement one mechanism for VMs and one mechanism for
+containers.
+
+Next, methods for extracting the guest and firewall configuration and the
+implementations of the restore mechanism via a pair of methods: an init method,
+for making the data available to Proxmox VE and a cleanup method that is called
+after restore.
+
+2.1. Restore Mechanisms
+
+VM mechanism 'qemu-img':
+
+The backup provider gives a path to the disk image that will be restored. The
+path needs to be something 'qemu-img' can deal with, e.g. can also be an NBD URI
+or similar.
+
+Container mechanism 'directory':
+
+The backup provider gives the path to a directory with the full filesystem
+structure of the container.
+
+Container mechanism 'tar':
+
+The backup provider gives the path to a (potentially compressed) tar archive
+with the full filesystem structure of the container.
+
+=head1 METHODS
+
+=cut
+
+# plugin methods
+
+=pod
+
+=over
+
+=item C<new>
+
+The constructor. Returns a blessed instance of the backup provider class.
+
+Parameters:
+
+=over
+
+=item C<$storage_plugin>
+
+The associated storage plugin class.
+
+=item C<$scfg>
+
+The storage configuration of the associated storage.
+
+=item C<$storeid>
+
+The storage ID of the associated storage.
+
+=item C<$log_function>
+
+The function signature is C<$log_function($log_level, $message)>. This log
+function can be used to write to the backup task log in Proxmox VE.
+
+=over
+
+=item C<$log_level>
+
+Either C<info>, C<warn> or C<err> for informational messages, warnings or error
+messages.
+
+=item C<$message>
+
+The message to be printed.
+
+=back
+
+=back
+
+=back
+
+=cut
+sub new {
+    my ($class, $storage_plugin, $scfg, $storeid, $log_function) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<provider_name>
+
+Returns the name of the backup provider. It will be printed in some log lines.
+
+=back
+
+=cut
+sub provider_name {
+    my ($self) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<job_init>
+
+Called when the job is started. Can be used to check the backup server
+availability and prepare for the upcoming backup tasks of individual guests. For
+example, to establish a connection to be used during C<backup_container()> or
+C<backup_vm()>.
+
+Parameters:
+
+=over
+
+=item C<$start_time>
+
+Unix time-stamp of when the job started.
+
+=back
+
+=back
+
+=cut
+sub job_init {
+    my ($self, $start_time) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<job_cleanup>
+
+Called when the job is finished to allow for any potential cleanup related to
+the backup server. Called in both, success and failure scenarios.
+
+=back
+
+=cut
+sub job_cleanup {
+    my ($self) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_init>
+
+Called before the backup of the given guest is made. The archive name is
+determined for the backup task and returned to the caller via a hash reference:
+
+    my $res = $backup_provider->backup_init($vmid, $vmtype, $start_time);
+    my $archive_name = $res->{'archive-name'};
+
+Use C<$self> to remember it for the C<backup_container()> or C<backup_vm()>
+method that will be called later.
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$vmtype>
+
+The type of the guest being backed up. Currently, either C<qemu> or C<lxc>.
+
+=item C<$start_time>
+
+Unix time-stamp of when the guest backup started.
+
+=back
+
+=back
+
+=cut
+sub backup_init {
+    my ($self, $vmid, $vmtype, $start_time) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_cleanup>
+
+Called when the guest backup is finished. Called in both, success and failure
+scenarios. In the success case, statistics about the task after completion of
+the backup are returned via a hash reference. Currently, only the archive size
+is part of the result:
+
+    my $res = $backup_provider->backup_cleanup($vmid, $vmtype, $success, $info);
+    my $stats = $res->{stats};
+    my $archive_size = $stats->{'archive-size'};
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$vmtype>
+
+The type of the guest being backed up. Currently, either C<qemu> or C<lxc>.
+Might be C<undef> in phase C<abort> for certain error scenarios.
+
+=item C<$success>
+
+Boolean indicating whether the job was successful or not. Success means that all
+individual guest backups were successful.
+
+=item C<$info>
+
+A hash reference with optional information. Currently, the error message in case
+of a failure.
+
+=over
+
+=item C<< $info->{error} >>
+
+Present if there was a failure. The error message indicating the failure.
+
+=back
+
+=back
+
+=back
+
+=cut
+sub backup_cleanup {
+    my ($self, $vmid, $vmtype, $success, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_container_prepare>
+
+Called right before C<backup_container()> is called. The method
+C<backup_container()> is called as the ID-mapped root user of the container, so
+as a potentially unprivileged user. The hook is still called as a privileged
+user to allow for the necessary preparation.
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$info>
+
+The same information that's passed along to C<backup_container()>, see the
+description there.
+
+=back
+
+=back
+
+=cut
+sub backup_container_prepare {
+    my ($self, $vmid, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_get_mechanism>
+
+Tell the caller what mechanism to use for backing up the guest. The backup
+method for the guest, i.e. C<backup_vm> for guest type C<qemu> or
+C<backup_container> for guest type C<lxc>, will later be called with
+mechanism-specific information. See those methods for more information. Returns
+C<($mechanism, $bitmap_id)>:
+
+=over
+
+=item C<$mechanism>
+
+Currently C<nbd> and C<file-handle> for guest type C<qemu> and C<directory> for
+guest type C<lxc> are possible. If there is no support for one of the guest
+types, the method should either C<die> or return C<undef>.
+
+=item C<$bitmap_id>
+
+If the backup provider supports backing up with a bitmap, the ID of the bitmap
+to use. Return C<undef> otherwise. Re-use the same ID multiple times for
+incremental backup.
+
+=back
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$vmtype>
+
+The type of the guest being backed up. Currently, either C<qemu> or C<lxc>.
+
+=back
+
+=back
+
+=cut
+sub backup_get_mechanism {
+    my ($self, $vmid, $vmtype) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_handle_log_file>
+
+Handle the backup's log file which contains the task log for the backup. For
+example, a provider might want to upload a copy to the backup server.
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$filename>
+
+Path to the file with the backup log.
+
+=back
+
+=back
+
+=cut
+sub backup_handle_log_file {
+    my ($self, $vmid, $filename) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_vm>
+
+Used when the guest type is C<qemu>. Back up the virtual machine's configuration
+and volumes that were made available according to the mechanism returned by
+C<backup_get_mechanism>. Returns when done backing up. Ideally, the method
+should log the progress during backup.
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$guest_config>
+
+The guest configuration as raw data.
+
+=item C<$volumes>
+
+Hash reference with information about the VM's volumes. Some parameters are
+mechanism-specific.
+
+=over
+
+=item C<< $volumes->{$device_name} >>
+
+Hash reference with information about the VM volume associated to the device
+C<$device_name>. The device name needs to be remembered for restoring. The
+device name is also the name of the NBD export when the C<nbd> mechanism is
+used.
+
+=item C<< $volumes->{$device_name}->{size} >>
+
+Size of the volume in bytes.
+
+=item C<< $volumes->{$device_name}->{'bitmap-mode'} >>
+
+How a bitmap is used for the current volume.
+
+=over
+
+=item C<none>
+
+No bitmap is used.
+
+=item C<new>
+
+A bitmap has been newly created on the volume.
+
+=item C<reuse>
+
+The bitmap with the same ID as requested is being re-used.
+
+=back
+
+=back
+
+Mechansims-specific parameters for mechanism:
+
+=over
+
+=item C<file-handle>
+
+=over
+
+=item C<< $volumes->{$device_name}->{'file-handle'} >>
+
+File handle the backup data can be read from. Discards should be issued via the
+C<PVE::Storage::Common::deallocate()> function for ranges that already have been
+backed-up successfully to reduce space usage on the source-side.
+
+=item C<< $volumes->{$device_name}->{'next-dirty-region'} >>
+
+A function that will return the offset and length of the next dirty region as a
+two-element list. After the last dirty region, it will return C<undef>. If no
+bitmap is used, it will return C<(0, $size)> and then C<undef>. If a bitmap is
+used, these are the dirty regions according to the bitmap.
+
+=back
+
+=item C<nbd>
+
+=over
+
+=item C<< $volumes->{$device_name}->{'nbd-path'} >>
+
+The path to the Unix socket providing the NBD export with the backup data and,
+if a bitmap is used, bitmap data. Discards should be issued after reading the
+data to reduce space usage on the source-side.
+
+=item C<< $volumes->{$device_name}->{'bitmap-name'} >>
+
+The name of the bitmap in case a bitmap is used.
+
+=back
+
+=back
+
+=item C<$info>
+
+A hash reference containing optional parameters.
+
+Optional parameters:
+
+=over
+
+=item C<< $info->{'bandwidth-limit'} >>
+
+The requested bandwith limit. The value is in bytes/second. The backup provider
+is expected to honor this rate limit for IO on the backup source and network
+traffic. A value of C<0>, C<undef> or if there is no such key in the hash all
+mean that there is no limit.
+
+=item C<< $info->{'firewall-config'} >>
+
+Present if the firewall configuration exists. The guest's firewall
+configuration as raw data.
+
+=back
+
+=back
+
+=back
+
+=cut
+sub backup_vm {
+    my ($self, $vmid, $guest_config, $volumes, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<backup_container>
+
+Used when the guest type is C<lxc>. Back up the container filesystem structure
+that is made available for the mechanism returned by C<backup_get_mechanism>.
+Returns when done backing up. Ideally, the method should log the progress during
+backup.
+
+Note that this function is executed as the ID-mapped root user of the container,
+so a potentially unprivileged user. The ID is passed along as part of C<$info>.
+Use the C<backup_container_prepare()> method for preparation. For example, to
+make credentials available to the potentially unprivileged user.
+
+Parameters:
+
+=over
+
+=item C<$vmid>
+
+The ID of the guest being backed up.
+
+=item C<$guest_config>
+
+Guest configuration as raw data.
+
+=item C<$exclude_patterns>
+
+A list of glob patterns of files and directories to be excluded. C<**> is used
+to match current directory and subdirectories. See also the following (note
+that PBS implements more than required here, like explicit inclusion when
+starting with a C<!>):
+L<vzdump documentation|https://pve.proxmox.com/pve-docs/chapter-vzdump.html#_file_exclusions>
+and
+L<PBS documentation|https://pbs.proxmox.com/docs/backup-client.html#excluding-files-directories-from-a-backup>
+
+=item C<$info>
+
+A hash reference containing optional and mechanism-specific parameters.
+
+Optional parameters:
+
+=over
+
+=item C<< $info->{'bandwidth-limit'} >>
+
+The requested bandwith limit. The value is in bytes/second. The backup provider
+is expected to honor this rate limit for IO on the backup source and network
+traffic. A value of C<0>, C<undef> or if there is no such key in the hash all
+mean that there is no limit.
+
+=item C<< $info->{'firewall-config'} >>
+
+Present if the firewall configuration exists. The guest's firewall
+configuration as raw data.
+
+=back
+
+Mechansims-specific parameters for mechanism:
+
+=over
+
+=item C<directory>
+
+=over
+
+=item C<< $info->{directory} >>
+
+Path to the directory with the container's file system structure.
+
+=item C<< $info->{sources} >>
+
+List of paths (for separate mount points, including "." for the root) inside the
+directory to be backed up.
+
+=item C<< $info->{'backup-user-id'} >>
+
+The user ID of the ID-mapped root user of the container. For example, C<100000>
+for unprivileged containers by default.
+
+=back
+
+=back
+
+=back
+
+=back
+
+=cut
+sub backup_container {
+    my ($self, $vmid, $guest_config, $exclude_patterns, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_get_mechanism>
+
+Tell the caller what mechanism to use for restoring the guest. The restore
+methods for the guest, i.e. C<restore_qemu_img_init> and
+C<restore_qemu_img_cleanup> for guest type C<qemu>, or C<restore_container_init>
+and C<restore_container_cleanup> for guest type C<lxc> will be called with
+mechanism-specific information and their return value might also depend on the
+mechanism. See those methods for more information. Returns
+C<($mechanism, $vmtype)>:
+
+=over
+
+=item C<$mechanism>
+
+Currently, C<'qemu-img'> for guest type C<'qemu'> and either C<'tar'> or
+C<'directory'> for type C<'lxc'> are possible.
+
+=item C<$vmtype>
+
+Either C<qemu> or C<lxc> depending on what type the guest in the backed-up
+archive is.
+
+=back
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=back
+
+=back
+
+=cut
+sub restore_get_mechanism {
+    my ($self, $volname) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<archive_get_guest_config>
+
+Extract the guest configuration from the given backup. Returns the raw contents
+of the backed-up configuration file. Note that this method is called
+independently from C<restore_container_init()> or C<restore_vm_init()>.
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=back
+
+=back
+
+=cut
+sub archive_get_guest_config {
+    my ($self, $volname) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<archive_get_firewall_config>
+
+Extract the guest's firewall configuration from the given backup. Returns the
+raw contents of the backed-up configuration file. Returns C<undef> if there is
+no firewall config in the archive, C<die> if the configuration can't be
+extracted. Note that this method is called independently from
+C<restore_container_init()> or C<restore_vm_init()>.
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=back
+
+=back
+
+=cut
+sub archive_get_firewall_config {
+    my ($self, $volname) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_vm_init>
+
+Prepare a VM archive for restore. Returns the basic information about the
+volumes in the backup as a hash reference with the following structure:
+
+    {
+	$device_nameA => { size => $sizeA },
+	$device_nameB => { size => $sizeB },
+	...
+    }
+
+=over
+
+=item C<$device_name>
+
+The device name that was given as an argument to the backup routine when the
+backup was created.
+
+=item C<$size>
+
+The virtual size of the VM volume that was backed up. A volume with this size is
+created for the restore operation. In particular, for the C<qemu-img> mechanism,
+this should be the size of the block device referenced by the C<qemu-img-path>
+returned by C<restore_vm_volume>.
+
+=back
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=back
+
+=back
+
+=cut
+sub restore_vm_init {
+    my ($self, $volname) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_vm_cleanup>
+
+For VM backups, clean up after the restore. Called in both, success and
+failure scenarios.
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=back
+
+=back
+
+=cut
+sub restore_vm_cleanup {
+    my ($self, $volname) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_vm_volume_init>
+
+Prepare a VM volume in the archive for restore. Returns a hash reference with
+the mechanism-specific information for the restore:
+
+=over
+
+=item C<qemu-img>
+
+    { 'qemu-img-path' => $path }
+
+The volume will be restored using the C<qemu-img convert> command.
+
+=over
+
+=item C<$path>
+
+A path to the volume that C<qemu-img> can use as a source for the
+C<qemu-img convert> command. For example, the path could also be an NBD URI. The
+image contents are interpreted as being in C<raw> format and copied verbatim.
+Other formats like C<qcow2> will not be detected currently.
+
+=back
+
+=back
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=item C<$device_name>
+
+The device name associated to the volume that should be prepared for the
+restore. Same as the argument to the backup routine when the backup was created.
+
+=item C<$info>
+
+A hash reference with optional and mechanism-specific parameters. Currently
+empty.
+
+=back
+
+=back
+
+=cut
+sub restore_vm_volume_init {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_vm_volume_cleanup>
+
+For VM backups, clean up after the restore of a given volume. Called in both,
+success and failure scenarios.
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=item C<$device_name>
+
+The device name associated to the volume that should be prepared for the
+restore. Same as the argument to the backup routine when the backup was created.
+
+=item C<$info>
+
+A hash reference with optional and mechanism-specific parameters. Currently
+empty.
+
+=back
+
+=back
+
+=cut
+sub restore_vm_volume_cleanup {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_container_init>
+
+Prepare a container archive for restore. Returns a hash reference with the
+mechanism-specific information for the restore:
+
+=over
+
+=item C<tar>
+
+    { 'tar-path' => $path }
+
+The archive will be restored via the C<tar> command.
+
+=over
+
+=item C<$path>
+
+The path to the tar archive containing the full filesystem structure of the
+container.
+
+=back
+
+=item C<directory>
+
+    { 'archive-directory' => $path }
+
+The archive will be restored via C<rsync> from a directory containing the full
+filesystem structure of the container.
+
+=over
+
+=item C<$path>
+
+The path to the directory containing the full filesystem structure of the
+container.
+
+=back
+
+=back
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=item C<$info>
+
+A hash reference with optional and mechanism-specific parameters. Currently
+empty.
+
+=back
+
+=back
+
+=cut
+sub restore_container_init {
+    my ($self, $volname, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+=pod
+
+=over
+
+=item C<restore_container_cleanup>
+
+For container backups, clean up after the restore. Called in both, success and
+failure scenarios.
+
+Parameters:
+
+=over
+
+=item C<$volname>
+
+The volume ID of the archive being restored.
+
+=item C<$info>
+
+A hash reference with optional and mechanism-specific parameters. Currently
+empty.
+
+=back
+
+=back
+
+=cut
+sub restore_container_cleanup {
+    my ($self, $volname, $info) = @_;
+
+    die "implement me in subclass";
+}
+
+1;
diff --git a/src/PVE/BackupProvider/Plugin/Makefile b/src/PVE/BackupProvider/Plugin/Makefile
new file mode 100644
index 0000000..bbd7431
--- /dev/null
+++ b/src/PVE/BackupProvider/Plugin/Makefile
@@ -0,0 +1,5 @@
+SOURCES = Base.pm
+
+.PHONY: install
+install:
+	for i in ${SOURCES}; do install -D -m 0644 $$i ${DESTDIR}${PERLDIR}/PVE/BackupProvider/Plugin/$$i; done
diff --git a/src/PVE/Makefile b/src/PVE/Makefile
index 0af3081..9e9f6aa 100644
--- a/src/PVE/Makefile
+++ b/src/PVE/Makefile
@@ -9,6 +9,7 @@ install:
 	make -C Storage install
 	make -C GuestImport install
 	make -C API2 install
+	make -C BackupProvider install
 	make -C CLI install
 
 .PHONY: test
diff --git a/src/PVE/Storage.pm b/src/PVE/Storage.pm
index 8cbfb4f..7fd97b7 100755
--- a/src/PVE/Storage.pm
+++ b/src/PVE/Storage.pm
@@ -2027,6 +2027,14 @@ sub volume_export_start {
     PVE::Tools::run_command($cmds, %$run_command_params);
 }
 
+sub new_backup_provider {
+    my ($cfg, $storeid, $log_function) = @_;
+
+    my $scfg = storage_config($cfg, $storeid);
+    my $plugin = PVE::Storage::Plugin->lookup($scfg->{type});
+    return $plugin->new_backup_provider($scfg, $storeid, $log_function);
+}
+
 # bash completion helper
 
 sub complete_storage {
diff --git a/src/PVE/Storage/Plugin.pm b/src/PVE/Storage/Plugin.pm
index 80daeea..df2ddc5 100644
--- a/src/PVE/Storage/Plugin.pm
+++ b/src/PVE/Storage/Plugin.pm
@@ -1868,6 +1868,21 @@ sub rename_volume {
     return "${storeid}:${base}${target_vmid}/${target_volname}";
 }
 
+# Used by storage plugins for external backup providers. See PVE::BackupProvider::Plugin for the API
+# the provider needs to implement.
+#
+# $scfg - the storage configuration
+# $storeid - the storage ID
+# $log_function($log_level, $message) - this log function can be used to write to the backup task
+#   log in Proxmox VE. $log_level is 'info', 'warn' or 'err', $message is the message to be printed.
+#
+# Returns a blessed reference to the backup provider class.
+sub new_backup_provider {
+    my ($class, $scfg, $storeid, $log_function) = @_;
+
+    die "implement me if enabling the feature 'backup-provider' in plugindata()->{features}\n";
+}
+
 sub config_aware_base_mkdir {
     my ($class, $scfg, $path) = @_;
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 13/37] config api/plugins: let plugins define sensitive properties themselves

[Fiona Ebner]

Hard-coding a list of sensitive properties means that custom plugins
cannot define their own sensitive properties for the on_add/on_update
hooks.

Have plugins declare the list of their sensitive properties in the
plugin data. For backwards compatibility, return the previously
hard-coded list if no such declaration is present.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/API2/Storage/Config.pm       |  4 ++--
 src/PVE/Storage/BTRFSPlugin.pm       |  1 +
 src/PVE/Storage/CIFSPlugin.pm        |  1 +
 src/PVE/Storage/CephFSPlugin.pm      |  1 +
 src/PVE/Storage/DirPlugin.pm         |  1 +
 src/PVE/Storage/ESXiPlugin.pm        |  1 +
 src/PVE/Storage/GlusterfsPlugin.pm   |  1 +
 src/PVE/Storage/ISCSIDirectPlugin.pm |  1 +
 src/PVE/Storage/ISCSIPlugin.pm       |  1 +
 src/PVE/Storage/LVMPlugin.pm         |  1 +
 src/PVE/Storage/LvmThinPlugin.pm     |  1 +
 src/PVE/Storage/NFSPlugin.pm         |  1 +
 src/PVE/Storage/PBSPlugin.pm         |  5 +++++
 src/PVE/Storage/Plugin.pm            | 12 ++++++++++++
 src/PVE/Storage/RBDPlugin.pm         |  1 +
 src/PVE/Storage/ZFSPlugin.pm         |  1 +
 src/PVE/Storage/ZFSPoolPlugin.pm     |  1 +
 17 files changed, 33 insertions(+), 2 deletions(-)

diff --git a/src/PVE/API2/Storage/Config.pm b/src/PVE/API2/Storage/Config.pm
index e04b6ab..7facc62 100755
--- a/src/PVE/API2/Storage/Config.pm
+++ b/src/PVE/API2/Storage/Config.pm
@@ -190,8 +190,6 @@ __PACKAGE__->register_method ({
 	return &$api_storage_config($cfg, $param->{storage});
     }});
 
-my $sensitive_params = [qw(password encryption-key master-pubkey keyring)];
-
 __PACKAGE__->register_method ({
     name => 'create',
     protected => 1,
@@ -239,6 +237,7 @@ __PACKAGE__->register_method ({
 	# fix me in section config create never need an empty entity.
 	delete $param->{nodes} if !$param->{nodes};
 
+	my $sensitive_params = PVE::Storage::Plugin::sensitive_properties($type);
 	my $sensitive = extract_sensitive_params($param, $sensitive_params, []);
 
 	my $plugin = PVE::Storage::Plugin->lookup($type);
@@ -344,6 +343,7 @@ __PACKAGE__->register_method ({
 	    my $scfg = PVE::Storage::storage_config($cfg, $storeid);
 	    $type = $scfg->{type};
 
+	    my $sensitive_params = PVE::Storage::Plugin::sensitive_properties($type);
 	    my $sensitive = extract_sensitive_params($param, $sensitive_params, $delete);
 
 	    my $plugin = PVE::Storage::Plugin->lookup($type);
diff --git a/src/PVE/Storage/BTRFSPlugin.pm b/src/PVE/Storage/BTRFSPlugin.pm
index 1966b6f..5ed910d 100644
--- a/src/PVE/Storage/BTRFSPlugin.pm
+++ b/src/PVE/Storage/BTRFSPlugin.pm
@@ -45,6 +45,7 @@ sub plugindata {
 	    { images => 1, rootdir => 1 },
 	],
 	format => [ { raw => 1, subvol => 1 }, 'raw', ],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/CIFSPlugin.pm b/src/PVE/Storage/CIFSPlugin.pm
index 475065a..f47861e 100644
--- a/src/PVE/Storage/CIFSPlugin.pm
+++ b/src/PVE/Storage/CIFSPlugin.pm
@@ -101,6 +101,7 @@ sub plugindata {
 	content => [ { images => 1, rootdir => 1, vztmpl => 1, iso => 1,
 		   backup => 1, snippets => 1, import => 1}, { images => 1 }],
 	format => [ { raw => 1, qcow2 => 1, vmdk => 1 } , 'raw' ],
+	'sensitive-properties' => { password => 1 },
     };
 }
 
diff --git a/src/PVE/Storage/CephFSPlugin.pm b/src/PVE/Storage/CephFSPlugin.pm
index 36c64ea..73edecb 100644
--- a/src/PVE/Storage/CephFSPlugin.pm
+++ b/src/PVE/Storage/CephFSPlugin.pm
@@ -118,6 +118,7 @@ sub plugindata {
     return {
 	content => [ { vztmpl => 1, iso => 1, backup => 1, snippets => 1, import => 1 },
 		     { backup => 1 }],
+	'sensitive-properties' => { keyring => 1 },
     };
 }
 
diff --git a/src/PVE/Storage/DirPlugin.pm b/src/PVE/Storage/DirPlugin.pm
index fb23e0a..532701b 100644
--- a/src/PVE/Storage/DirPlugin.pm
+++ b/src/PVE/Storage/DirPlugin.pm
@@ -26,6 +26,7 @@ sub plugindata {
 	content => [ { images => 1, rootdir => 1, vztmpl => 1, iso => 1, backup => 1, snippets => 1, none => 1, import => 1 },
 		     { images => 1,  rootdir => 1 }],
 	format => [ { raw => 1, qcow2 => 1, vmdk => 1, subvol => 1 } , 'raw' ],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/ESXiPlugin.pm b/src/PVE/Storage/ESXiPlugin.pm
index c8412c4..6131c51 100644
--- a/src/PVE/Storage/ESXiPlugin.pm
+++ b/src/PVE/Storage/ESXiPlugin.pm
@@ -31,6 +31,7 @@ sub plugindata {
     return {
 	content => [ { import => 1 }, { import => 1 }],
 	format => [ { raw => 1, qcow2 => 1, vmdk => 1 } , 'raw' ],
+	'sensitive-properties' => { password => 1 },
     };
 }
 
diff --git a/src/PVE/Storage/GlusterfsPlugin.pm b/src/PVE/Storage/GlusterfsPlugin.pm
index 9d17180..18493cb 100644
--- a/src/PVE/Storage/GlusterfsPlugin.pm
+++ b/src/PVE/Storage/GlusterfsPlugin.pm
@@ -100,6 +100,7 @@ sub plugindata {
 	content => [ { images => 1, vztmpl => 1, iso => 1, backup => 1, snippets => 1, import => 1},
 		     { images => 1 }],
 	format => [ { raw => 1, qcow2 => 1, vmdk => 1 } , 'raw' ],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/ISCSIDirectPlugin.pm b/src/PVE/Storage/ISCSIDirectPlugin.pm
index 60bc94e..829e0c4 100644
--- a/src/PVE/Storage/ISCSIDirectPlugin.pm
+++ b/src/PVE/Storage/ISCSIDirectPlugin.pm
@@ -60,6 +60,7 @@ sub plugindata {
     return {
 	content => [ {images => 1, none => 1}, { images => 1 }],
 	select_existing => 1,
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/ISCSIPlugin.pm b/src/PVE/Storage/ISCSIPlugin.pm
index 2c608a4..39cb4a7 100644
--- a/src/PVE/Storage/ISCSIPlugin.pm
+++ b/src/PVE/Storage/ISCSIPlugin.pm
@@ -305,6 +305,7 @@ sub plugindata {
     return {
 	content => [ {images => 1, none => 1}, { images => 1 }],
 	select_existing => 1,
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/LVMPlugin.pm b/src/PVE/Storage/LVMPlugin.pm
index 38f7fa1..2ebec88 100644
--- a/src/PVE/Storage/LVMPlugin.pm
+++ b/src/PVE/Storage/LVMPlugin.pm
@@ -218,6 +218,7 @@ sub type {
 sub plugindata {
     return {
 	content => [ {images => 1, rootdir => 1}, { images => 1 }],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/LvmThinPlugin.pm b/src/PVE/Storage/LvmThinPlugin.pm
index 4b23623..49a4dcb 100644
--- a/src/PVE/Storage/LvmThinPlugin.pm
+++ b/src/PVE/Storage/LvmThinPlugin.pm
@@ -31,6 +31,7 @@ sub type {
 sub plugindata {
     return {
 	content => [ {images => 1, rootdir => 1}, { images => 1, rootdir => 1}],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/NFSPlugin.pm b/src/PVE/Storage/NFSPlugin.pm
index 72e9c6d..cb2ae18 100644
--- a/src/PVE/Storage/NFSPlugin.pm
+++ b/src/PVE/Storage/NFSPlugin.pm
@@ -56,6 +56,7 @@ sub plugindata {
 	content => [ { images => 1, rootdir => 1, vztmpl => 1, iso => 1, backup => 1, snippets => 1, import => 1 },
 		     { images => 1 }],
 	format => [ { raw => 1, qcow2 => 1, vmdk => 1 } , 'raw' ],
+	'sensitive-properties' => {},
     };
 }   
 
diff --git a/src/PVE/Storage/PBSPlugin.pm b/src/PVE/Storage/PBSPlugin.pm
index 0808bcc..9f75794 100644
--- a/src/PVE/Storage/PBSPlugin.pm
+++ b/src/PVE/Storage/PBSPlugin.pm
@@ -30,6 +30,11 @@ sub type {
 sub plugindata {
     return {
 	content => [ {backup => 1, none => 1}, { backup => 1 }],
+	'sensitive-properties' => {
+	    'encryption-key' => 1,
+	    'master-pubkey' => 1,
+	    password => 1,
+	},
     };
 }
 
diff --git a/src/PVE/Storage/Plugin.pm b/src/PVE/Storage/Plugin.pm
index df2ddc5..0d9558c 100644
--- a/src/PVE/Storage/Plugin.pm
+++ b/src/PVE/Storage/Plugin.pm
@@ -246,6 +246,18 @@ sub dirs_hash_to_string {
     return join(',', map { "$_=$hash->{$_}" } sort keys %$hash);
 }
 
+sub sensitive_properties {
+    my ($type) = @_;
+
+    my $data = $defaultData->{plugindata}->{$type};
+    if (my $sensitive_properties = $data->{'sensitive-properties'}) {
+	return [sort keys $sensitive_properties->%*];
+    }
+
+    # For backwards compatibility. This list was hardcoded in the API module previously.
+    return [qw(encryption-key keyring master-pubkey password)];
+}
+
 sub storage_has_feature {
     my ($type, $feature) = @_;
 
diff --git a/src/PVE/Storage/RBDPlugin.pm b/src/PVE/Storage/RBDPlugin.pm
index 42eefc6..c78db00 100644
--- a/src/PVE/Storage/RBDPlugin.pm
+++ b/src/PVE/Storage/RBDPlugin.pm
@@ -380,6 +380,7 @@ sub type {
 sub plugindata {
     return {
 	content => [ {images => 1, rootdir => 1}, { images => 1 }],
+	'sensitive-properties' => { keyring => 1 },
     };
 }
 
diff --git a/src/PVE/Storage/ZFSPlugin.pm b/src/PVE/Storage/ZFSPlugin.pm
index d4dc2a4..94cb11f 100644
--- a/src/PVE/Storage/ZFSPlugin.pm
+++ b/src/PVE/Storage/ZFSPlugin.pm
@@ -175,6 +175,7 @@ sub type {
 sub plugindata {
     return {
 	content => [ {images => 1}, { images => 1 }],
+	'sensitive-properties' => {},
     };
 }
 
diff --git a/src/PVE/Storage/ZFSPoolPlugin.pm b/src/PVE/Storage/ZFSPoolPlugin.pm
index 3669fe1..26fb0a4 100644
--- a/src/PVE/Storage/ZFSPoolPlugin.pm
+++ b/src/PVE/Storage/ZFSPoolPlugin.pm
@@ -22,6 +22,7 @@ sub plugindata {
     return {
 	content => [ {images => 1, rootdir => 1}, {images => 1 , rootdir => 1}],
 	format => [ { raw => 1, subvol => 1 } , 'raw' ],
+	'sensitive-properties' => {},
     };
 }
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 14/37] plugin api: bump api version and age

[Fiona Ebner]

Changes for version 11:

* Allow declaring storage features via plugin data.
* Introduce new_backup_provider() plugin method.
* Allow declaring sensitive properties via plugin data.

See the api changelog file for details.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 ApiChangeLog       | 32 ++++++++++++++++++++++++++++++++
 src/PVE/Storage.pm |  4 ++--
 2 files changed, 34 insertions(+), 2 deletions(-)

diff --git a/ApiChangeLog b/ApiChangeLog
index 98b5893..987da54 100644
--- a/ApiChangeLog
+++ b/ApiChangeLog
@@ -6,6 +6,38 @@ without breaking anything unaware of it.)
 
 Future changes should be documented in here.
 
+##  Version 11:
+
+* Allow declaring storage features via plugin data
+
+  A new `storage_has_feature()` helper function was added that checks a storage plugin's features.
+  Plugins can indicate support for certain features in their `plugindata`. The first such feature is
+  `backup-provider`, see below for more details. To declare support for this feature, return
+  `features => { 'backup-provider' => 1 }` as part of the plugin data.
+
+* Introduce new_backup_provider() plugin method
+
+  Proxmox VE now supports a `Backup Provider API` that can be used to implement custom backup
+  solutions tightly integrated in the Proxmox VE stack. See the `PVE::BackupProvider::Plugin::Base`
+  module for detailed documentation. A backup provider also needs to implement an associated storage
+  plugin for user-facing integration in Proxmox VE. Such a plugin needds to opt-in to the
+  `backup-provider` feature (see above) and implement the new_backup_provider() method, returning a
+  blessed reference to the backup provider class. The rest of the plugin methods, e.g. listing
+  content, providing usage information, etc., follow the same API as usual.
+
+* Allow declaring sensitive properties via plugin data
+
+  A new `sensitive_properties()` helper function was added to get the list of sensitive properties
+  a plugin uses via the plugin's `plugindata`. The sensitive properties are passed separately from
+  other properties to the `on_add_hook()` and `on_update_hook()` methods and should not be written
+  to the storage configuration file directly, but stored in the more restricted
+  `/etc/pve/priv/storage` directory on the Proxmox Cluster File System. For example, to declare that
+  a `ssh-private-key` property used by the plugin is sensitive, return
+  `'sensitive-properties' => { 'ssh-private-key' => 1 }` as part of the plugin data. The list of
+  sensitive properties was hard-coded previously, as `encryption-key`, `keyring`, `master-pubkey`,
+  `password`. For backwards compatibility, this list is still used if a plugin doesn't declare its
+  own sensitive properties.
+
 ##  Version 10:
 
 * a new `rename_volume` method has been added
diff --git a/src/PVE/Storage.pm b/src/PVE/Storage.pm
index 7fd97b7..10a4abc 100755
--- a/src/PVE/Storage.pm
+++ b/src/PVE/Storage.pm
@@ -42,11 +42,11 @@ use PVE::Storage::BTRFSPlugin;
 use PVE::Storage::ESXiPlugin;
 
 # Storage API version. Increment it on changes in storage API interface.
-use constant APIVER => 10;
+use constant APIVER => 11;
 # Age is the number of versions we're backward compatible with.
 # This is like having 'current=APIVER' and age='APIAGE' in libtool,
 # see https://www.gnu.org/software/libtool/manual/html_node/Libtool-versioning.html
-use constant APIAGE => 1;
+use constant APIAGE => 2;
 
 our $KNOWN_EXPORT_FORMATS = ['raw+size', 'tar+size', 'qcow2+size', 'vmdk+size', 'zfs', 'btrfs'];
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH storage v6 15/37] extract backup config: delegate to backup provider for storages that support it

[Fiona Ebner]

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Adapt to renamed archive_get_guest_config() method.

 src/PVE/Storage.pm | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/src/PVE/Storage.pm b/src/PVE/Storage.pm
index 10a4abc..7174f0f 100755
--- a/src/PVE/Storage.pm
+++ b/src/PVE/Storage.pm
@@ -1759,6 +1759,17 @@ sub extract_vzdump_config {
 	    storage_check_enabled($cfg, $storeid);
 	    return PVE::Storage::PBSPlugin->extract_vzdump_config($scfg, $volname, $storeid);
 	}
+
+	if (storage_has_feature($cfg, $storeid, 'backup-provider')) {
+	    my $plugin = PVE::Storage::Plugin->lookup($scfg->{type});
+	    my $log_function = sub {
+		my ($log_level, $message) = @_;
+		my $prefix = $log_level eq 'err' ? 'ERROR' : uc($log_level);
+		print "$prefix: $message\n";
+	    };
+	    my $backup_provider = $plugin->new_backup_provider($scfg, $storeid, $log_function);
+	    return $backup_provider->archive_get_guest_config($volname, $storeid);
+	}
     }
 
     my $archive = abs_filesystem_path($cfg, $volid);
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [POC storage v6 16/37] add backup provider example

[Fiona Ebner]

The example uses a simple directory structure to save the backups,
grouped by guest ID. VM backups are saved as configuration files and
qcow2 images, with backing files when doing incremental backups.
Container backups are saved as configuration files and a tar file or
squashfs image (added to test the 'directory' restore mechanism).

Whether to use incremental VM backups and which backup mechanisms to
use can be configured in the storage configuration.

The 'nbdinfo' binary from the 'libnbd-bin' package is required for
backup mechanism 'nbd' for VM backups, the 'mksquashfs' binary from the
'squashfs-tools' package is required for backup mechanism 'squashfs' for
containers.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Remove superfluous $storeid parameter from methods. It's already
  part of $self.
* Rename restore_get_*_config() methods to archive_get_*_config().
* Replace job/backup hooks with dedicated methods.
* Remove backup_get_archive_name() method. The archive name is now
  returned as part of the backup_init() method.
* Supersede backup_get_task_size() method and return
  extensible-in-the-future stats as part of the backup_cleanup()
  method.

 .../BackupProvider/Plugin/DirectoryExample.pm | 778 ++++++++++++++++++
 src/PVE/BackupProvider/Plugin/Makefile        |   2 +-
 .../Custom/BackupProviderDirExamplePlugin.pm  | 308 +++++++
 src/PVE/Storage/Custom/Makefile               |   5 +
 src/PVE/Storage/Makefile                      |   1 +
 5 files changed, 1093 insertions(+), 1 deletion(-)
 create mode 100644 src/PVE/BackupProvider/Plugin/DirectoryExample.pm
 create mode 100644 src/PVE/Storage/Custom/BackupProviderDirExamplePlugin.pm
 create mode 100644 src/PVE/Storage/Custom/Makefile

diff --git a/src/PVE/BackupProvider/Plugin/DirectoryExample.pm b/src/PVE/BackupProvider/Plugin/DirectoryExample.pm
new file mode 100644
index 0000000..4d90c58
--- /dev/null
+++ b/src/PVE/BackupProvider/Plugin/DirectoryExample.pm
@@ -0,0 +1,778 @@
+package PVE::BackupProvider::Plugin::DirectoryExample;
+
+use strict;
+use warnings;
+
+use Fcntl qw(SEEK_SET);
+use File::Path qw(make_path remove_tree);
+use IO::File;
+use IPC::Open3;
+
+use PVE::Storage::Common;
+use PVE::Storage::Plugin;
+use PVE::Tools qw(file_get_contents file_read_firstline file_set_contents run_command);
+
+use base qw(PVE::BackupProvider::Plugin::Base);
+
+# Private helpers
+
+my sub log_info {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('info', $message);
+}
+
+my sub log_warning {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('warn', $message);
+}
+
+my sub log_error {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('err', $message);
+}
+
+# Try to use the same bitmap ID as last time for incremental backup if the storage is configured for
+# incremental VM backup. Need to start fresh if there is no previous ID or the associated backup
+# doesn't exist.
+my sub get_bitmap_id {
+    my ($self, $vmid, $vmtype) = @_;
+
+    return if $self->{'storage-plugin'}->get_vm_backup_mode($self->{scfg}) ne 'incremental';
+
+    my $previous_info_dir = "$self->{scfg}->{path}/$vmid/";
+
+    my $previous_info_file = "$previous_info_dir/previous-info";
+    my $info = file_read_firstline($previous_info_file) // '';
+    $self->{$vmid}->{'old-previous-info'} = $info;
+    my ($bitmap_id, $previous_backup_id) = $info =~ m/^(\d+)\s+(\d+)$/;
+    my $previous_backup_dir =
+	$previous_backup_id ? "$self->{scfg}->{path}/$vmid/$vmtype-$previous_backup_id" : undef;
+
+    if ($bitmap_id && -d $previous_backup_dir) {
+	$self->{$vmid}->{'previous-backup-dir'} = $previous_backup_dir;
+    } else {
+	# need to start fresh if there is no previous ID or the associated backup doesn't exist
+	$bitmap_id = $self->{$vmid}->{'backup-time'};
+    }
+
+    $self->{$vmid}->{'bitmap-id'} = $bitmap_id;
+    make_path($previous_info_dir);
+    die "unable to create directory $previous_info_dir\n" if !-d $previous_info_dir;
+    file_set_contents($previous_info_file, "$bitmap_id $self->{$vmid}->{'backup-time'}");
+
+    return $bitmap_id;
+}
+
+# NOTE: This is just for proof-of-concept testing! A backup provider plugin should either use the
+# 'nbd' backup mechansim and use the NBD protocol or use the 'file-handle' mechanism. There should
+# be no need to use /dev/nbdX nodes for proper plugins.
+my sub bind_next_free_dev_nbd_node {
+    my ($options) = @_;
+
+    # /dev/nbdX devices are reserved in a file. Those reservations expires after $expiretime.
+    # This avoids race conditions between allocation and use.
+
+    die "file '/sys/module/nbd' does not exist - 'nbd' kernel module not loaded?"
+	if !-e "/sys/module/nbd";
+
+    my $line = PVE::Tools::file_read_firstline("/sys/module/nbd/parameters/nbds_max")
+	or die "could not read 'nbds_max' parameter file for 'nbd' kernel module\n";
+    my ($nbds_max) = ($line =~ m/(\d+)/)
+	or die "could not determine 'nbds_max' parameter for 'nbd' kernel module\n";
+
+    my $filename = "/run/qemu-server/reserved-dev-nbd-nodes";
+
+    my $code = sub {
+	my $expiretime = 60;
+	my $ctime = time();
+
+	my $used = {};
+	my $latest = [0, 0];
+
+	if (my $fh = IO::File->new ($filename, "r")) {
+	    while (my $line = <$fh>) {
+		if ($line =~ m/^(\d+)\s(\d+)$/) {
+		    my ($n, $timestamp) = ($1, $2);
+
+		    $latest = [$n, $timestamp] if $latest->[1] <= $timestamp;
+
+		    if (($timestamp + $expiretime) > $ctime) {
+			$used->{$n} = $timestamp; # not expired
+		    }
+		}
+	    }
+	}
+
+	my $new_n;
+	for (my $count = 0; $count < $nbds_max; $count++) {
+	    my $n = ($latest->[0] + $count) % $nbds_max;
+	    my $block_device = "/dev/nbd${n}";
+	    next if $used->{$n}; # reserved
+	    next if !-e $block_device;
+
+	    my $st = File::stat::stat("/run/lock/qemu-nbd-nbd${n}");
+	    next if defined($st) && S_ISSOCK($st->mode) && $st->uid == 0; # in use
+
+	    # Used to avoid looping if there are other issues then the NBD node being in use
+	    my $socket_error = 0;
+	    eval {
+		my $errfunc = sub {
+		    my ($line) = @_;
+		    $socket_error = 1 if $line =~ m/^qemu-nbd: Failed to set NBD socket$/;
+		    log_warn($line);
+		};
+		run_command(["qemu-nbd", "-c", $block_device, $options->@*], errfunc => $errfunc);
+	    };
+	    if (my $err = $@) {
+		die $err if !$socket_error;
+		log_warn("unable to bind $block_device - trying next one");
+		next;
+	    }
+	    $used->{$n} = $ctime;
+	    $new_n = $n;
+	    last;
+	}
+
+	my $data = "";
+	$data .= "$_ $used->{$_}\n" for keys $used->%*;
+
+	PVE::Tools::file_set_contents($filename, $data);
+
+	return defined($new_n) ? "/dev/nbd${new_n}" : undef;
+    };
+
+    my $block_device =
+	PVE::Tools::lock_file('/run/lock/qemu-server/reserved-dev-nbd-nodes.lock', 10, $code);
+    die $@ if $@;
+
+    die "unable to find free /dev/nbdX block device node\n" if !$block_device;
+
+    return $block_device;
+}
+
+# Backup Provider API
+
+sub new {
+    my ($class, $storage_plugin, $scfg, $storeid, $log_function) = @_;
+
+    my $self = bless {
+	scfg => $scfg,
+	storeid => $storeid,
+	'storage-plugin' => $storage_plugin,
+	'log-function' => $log_function,
+    }, $class;
+
+    return $self;
+}
+
+sub provider_name {
+    my ($self) = @_;
+
+    return 'dir provider example';
+}
+
+# Hooks
+
+sub job_init {
+    my ($self, $start_time) = @_;
+
+    log_info($self, "job init called");
+
+    if (!-e '/sys/module/nbd/parameters') {
+	die "required 'nbd' kernel module not loaded - use 'modprobe nbd nbds_max=128' to load it"
+	    ." manually\n";
+    }
+
+    log_info($self, "backup provider initialized successfully for new job $start_time");
+
+    return;
+}
+
+sub job_cleanup {
+    my ($self) = @_;
+
+    log_info($self, "job cleanup called");
+
+    return;
+}
+
+sub backup_init {
+    my ($self, $vmid, $vmtype, $backup_time) = @_;
+
+    my $archive = "${vmid}/${vmtype}-${backup_time}";
+
+    log_info($self, "backup start hook called");
+
+    my $backup_dir = $self->{scfg}->{path} . "/" . $archive;
+
+    make_path($backup_dir);
+    die "unable to create directory $backup_dir\n" if !-d $backup_dir;
+
+    $self->{$vmid}->{'backup-time'} = $backup_time;
+    $self->{$vmid}->{'backup-dir'} = $backup_dir;
+
+    $self->{$vmid}->{archive} = $archive;
+    return { 'archive-name' => $archive };
+}
+
+sub backup_cleanup {
+    my ($self, $vmid, $vmtype, $success, $info) = @_;
+
+    if ($success) {
+	log_info($self, "backup cleanup called - success");
+	my $size = 0;
+	my $backup_dir = $self->{$vmid}->{'backup-dir'};
+	my @backup_files = glob("$backup_dir/*");
+	$size += -s $_ for @backup_files;
+	my $stats = { 'archive-size' => $size };
+	return { 'stats' => $stats };
+    } else {
+	log_info($self, "backup cleanup called - failure");
+
+	$self->{$vmid}->{failed} = 1;
+
+	if (my $dir = $self->{$vmid}->{'backup-dir'}) {
+	    eval { remove_tree($dir) };
+	    log_warning($self, "unable to clean up $dir - $@") if $@;
+	}
+
+	# Restore old previous-info so next attempt can re-use bitmap again
+	if (my $info = $self->{$vmid}->{'old-previous-info'}) {
+	    my $previous_info_dir = "$self->{scfg}->{path}/$vmid/";
+	    my $previous_info_file = "$previous_info_dir/previous-info";
+	    file_set_contents($previous_info_file, $info);
+	}
+    }
+}
+
+sub backup_container_prepare {
+    my ($self, $vmid, $info) = @_;
+
+    my $dir = $self->{$vmid}->{'backup-dir'};
+    chown($info->{'backup-user-id'}, -1, $dir) or die "unable to change owner for $dir\n";
+
+    return;
+}
+
+sub backup_get_mechanism {
+    my ($self, $vmid, $vmtype) = @_;
+
+    return ('directory', undef) if $vmtype eq 'lxc';
+
+    if ($vmtype eq 'qemu') {
+	my $backup_mechanism = $self->{'storage-plugin'}->get_vm_backup_mechanism($self->{scfg});
+	return ($backup_mechanism, get_bitmap_id($self, $vmid, $vmtype));
+    }
+
+    die "unsupported guest type '$vmtype'\n";
+}
+
+sub backup_handle_log_file {
+    my ($self, $vmid, $filename) = @_;
+
+    my $log_dir = $self->{$vmid}->{'backup-dir'};
+    if ($self->{$vmid}->{failed}) {
+	$log_dir .= ".failed";
+    }
+    make_path($log_dir);
+    die "unable to create directory $log_dir\n" if !-d $log_dir;
+
+    my $data = file_get_contents($filename);
+    my $target = "${log_dir}/backup.log";
+    file_set_contents($target, $data);
+}
+
+my sub backup_file {
+    my ($self, $vmid, $device_name, $size, $in_fh, $bitmap_mode, $next_dirty_region, $bandwidth_limit) = @_;
+
+    # TODO honor bandwidth_limit
+
+    my $previous_backup_dir = $self->{$vmid}->{'previous-backup-dir'};
+    my $incremental = $previous_backup_dir && $bitmap_mode eq 'reuse';
+    my $target = "$self->{$vmid}->{'backup-dir'}/${device_name}.qcow2";
+    my $target_base = $incremental ? "${previous_backup_dir}/${device_name}.qcow2" : undef;
+    my $create_cmd = ["qemu-img", "create", "-f", "qcow2", $target, $size];
+    push $create_cmd->@*, "-b", $target_base, "-F", "qcow2" if $target_base;
+    run_command($create_cmd);
+
+    my $nbd_node;
+    eval {
+	# allows to easily write to qcow2 target
+	$nbd_node = bind_next_free_dev_nbd_node([$target, '--format=qcow2']);
+	# FIXME use nbdfuse like in qemu-server rather than qemu-nbd. Seems like there is a race and
+	# sysseek() can fail with "Invalid argument" if done too early...
+	sleep 1;
+
+	my $block_size = 4 * 1024 * 1024; # 4 MiB
+
+	my $out_fh = IO::File->new($nbd_node, "r+")
+	    or die "unable to open NBD backup target - $!\n";
+
+	my $buffer = '';
+	my $skip_discard;
+
+	while (scalar((my $region_offset, my $region_length) = $next_dirty_region->())) {
+	    sysseek($in_fh, $region_offset, SEEK_SET)
+		// die "unable to seek '$region_offset' in NBD backup source - $!\n";
+	    sysseek($out_fh, $region_offset, SEEK_SET)
+		// die "unable to seek '$region_offset' in NBD backup target - $!\n";
+
+	    my $local_offset = 0; # within the region
+	    while ($local_offset < $region_length) {
+		my $remaining = $region_length - $local_offset;
+		my $request_size = $remaining < $block_size ? $remaining : $block_size;
+		my $offset = $region_offset + $local_offset;
+
+		my $read = sysread($in_fh, $buffer, $request_size);
+		die "failed to read from backup source - $!\n" if !defined($read);
+		die "premature EOF while reading backup source\n" if $read == 0;
+
+		my $written = 0;
+		while ($written < $read) {
+		    my $res = syswrite($out_fh, $buffer, $request_size - $written, $written);
+		    die "failed to write to backup target - $!\n" if !defined($res);
+		    die "unable to progress writing to backup target\n" if $res == 0;
+		    $written += $res;
+		}
+
+		if (!$skip_discard) {
+		    eval { PVE::Storage::Common::deallocate($in_fh, $offset, $request_size); };
+		    if (my $err = $@) {
+			# Just assume that if one request didn't work, others won't either.
+			log_warning(
+			    $self, "discard source failed (skipping further discards) - $err");
+			$skip_discard = 1;
+		     }
+		 }
+
+		$local_offset += $request_size;
+	    }
+	}
+	$out_fh->sync();
+    };
+    my $err = $@;
+
+    eval { run_command(['qemu-nbd', '-d', $nbd_node ]); };
+    log_warning($self, "unable to disconnect NBD backup target - $@") if $@;
+
+    die $err if $err;
+}
+
+my sub backup_nbd {
+    my ($self, $vmid, $device_name, $size, $nbd_path, $bitmap_mode, $bitmap_name, $bandwidth_limit) = @_;
+
+    # TODO honor bandwidth_limit
+
+    die "need 'nbdinfo' binary from package libnbd-bin\n" if !-e "/usr/bin/nbdinfo";
+
+    my $nbd_info_uri = "nbd+unix:///${device_name}?socket=${nbd_path}";
+    my $qemu_nbd_uri = "nbd:unix:${nbd_path}:exportname=${device_name}";
+
+    my $cpid;
+    my $error_fh;
+    my $next_dirty_region;
+
+    # If there is no dirty bitmap, it can be treated as if there's a full dirty one. The output of
+    # nbdinfo is a list of tuples with offset, length, type, description. The first bit of 'type' is
+    # set when the bitmap is dirty, see QEMU's docs/interop/nbd.txt
+    my $dirty_bitmap = [];
+    if ($bitmap_mode ne 'none') {
+	my $input = IO::File->new();
+	my $info = IO::File->new();
+	$error_fh = IO::File->new();
+	my $nbdinfo_cmd = ["nbdinfo", $nbd_info_uri, "--map=qemu:dirty-bitmap:${bitmap_name}"];
+	$cpid = open3($input, $info, $error_fh, $nbdinfo_cmd->@*)
+	    or die "failed to spawn nbdinfo child - $!\n";
+
+	$next_dirty_region = sub {
+	    my ($offset, $length, $type);
+	    do {
+		my $line = <$info>;
+		return if !$line;
+		die "unexpected output from nbdinfo - $line\n"
+		    if $line !~ m/^\s*(\d+)\s*(\d+)\s*(\d+)/; # also untaints
+		($offset, $length, $type) = ($1, $2, $3);
+	    } while (($type & 0x1) == 0); # not dirty
+	    return ($offset, $length);
+	};
+    } else {
+	my $done = 0;
+	$next_dirty_region = sub {
+	    return if $done;
+	    $done = 1;
+	    return (0, $size);
+	};
+    }
+
+    my $nbd_node;
+    eval {
+	$nbd_node = bind_next_free_dev_nbd_node([$qemu_nbd_uri, "--format=raw", "--discard=on"]);
+
+	my $in_fh = IO::File->new($nbd_node, 'r+')
+	    or die "unable to open NBD backup source '$nbd_node' - $!\n";
+
+	backup_file(
+	    $self,
+	    $vmid,
+	    $device_name,
+	    $size,
+	    $in_fh,
+	    $bitmap_mode,
+	    $next_dirty_region,
+	    $bandwidth_limit,
+	);
+    };
+    my $err = $@;
+
+    eval { run_command(["qemu-nbd", "-d", $nbd_node ]); };
+    log_warning($self, "unable to disconnect NBD backup source - $@") if $@;
+
+    if ($cpid) {
+	my $waited;
+	my $wait_limit = 5;
+	for ($waited = 0; $waited < $wait_limit && waitpid($cpid, POSIX::WNOHANG) == 0; $waited++) {
+	    kill 15, $cpid if $waited == 0;
+	    sleep 1;
+	}
+	if ($waited == $wait_limit) {
+	    kill 9, $cpid;
+	    sleep 1;
+	    log_warning($self, "unable to collect nbdinfo child process")
+		if waitpid($cpid, POSIX::WNOHANG) == 0;
+	}
+    }
+
+    die $err if $err;
+}
+
+my sub backup_vm_volume {
+    my ($self, $vmid, $device_name, $info, $bandwidth_limit) = @_;
+
+    my $backup_mechanism = $self->{'storage-plugin'}->get_vm_backup_mechanism($self->{scfg});
+
+    if ($backup_mechanism eq 'nbd') {
+	backup_nbd(
+	    $self,
+	    $vmid,
+	    $device_name,
+	    $info->{size},
+	    $info->{'nbd-path'},
+	    $info->{'bitmap-mode'},
+	    $info->{'bitmap-name'},
+	    $bandwidth_limit,
+	);
+    } elsif ($backup_mechanism eq 'file-handle') {
+	backup_file(
+	    $self,
+	    $vmid,
+	    $device_name,
+	    $info->{size},
+	    $info->{'file-handle'},
+	    $info->{'bitmap-mode'},
+	    $info->{'next-dirty-region'},
+	    $bandwidth_limit,
+	);
+    } else {
+	die "internal error - unknown VM backup mechansim '$backup_mechanism'\n";
+    }
+}
+
+sub backup_vm {
+    my ($self, $vmid, $guest_config, $volumes, $info) = @_;
+
+    my $target = "$self->{$vmid}->{'backup-dir'}/guest.conf";
+    file_set_contents($target, $guest_config);
+
+    if (my $firewall_config = $info->{'firewall-config'}) {
+	$target = "$self->{$vmid}->{'backup-dir'}/firewall.conf";
+	file_set_contents($target, $firewall_config);
+    }
+
+    for my $device_name (sort keys $volumes->%*) {
+	backup_vm_volume(
+	    $self, $vmid, $device_name, $volumes->{$device_name}, $info->{'bandwidth-limit'});
+    }
+}
+
+my sub backup_directory_tar {
+    my ($self, $vmid, $directory, $exclude_patterns, $sources, $bandwidth_limit) = @_;
+
+    # essentially copied from PVE/VZDump/LXC.pm' archive()
+
+    # copied from PVE::Storage::Plugin::COMMON_TAR_FLAGS
+    my @tar_flags = qw(
+	--one-file-system
+	-p --sparse --numeric-owner --acls
+	--xattrs --xattrs-include=user.* --xattrs-include=security.capability
+	--warning=no-file-ignored --warning=no-xattr-write
+    );
+
+    my $tar = ['tar', 'cpf', '-', '--totals', @tar_flags];
+
+    push @$tar, "--directory=$directory";
+
+    my @exclude_no_anchored = ();
+    my @exclude_anchored = ();
+    for my $pattern ($exclude_patterns->@*) {
+	if ($pattern !~ m|^/|) {
+	    push @exclude_no_anchored, $pattern;
+	} else {
+	    push @exclude_anchored, $pattern;
+	}
+    }
+
+    push @$tar, '--no-anchored';
+    push @$tar, '--exclude=lost+found';
+    push @$tar, map { "--exclude=$_" } @exclude_no_anchored;
+
+    push @$tar, '--anchored';
+    push @$tar, map { "--exclude=.$_" } @exclude_anchored;
+
+    push @$tar, $sources->@*;
+
+    my $cmd = [ $tar ];
+
+    push @$cmd, [ 'cstream', '-t', $bandwidth_limit * 1024 ] if $bandwidth_limit;
+
+    my $target = "$self->{$vmid}->{'backup-dir'}/archive.tar";
+    push @{$cmd->[-1]}, \(">" . PVE::Tools::shellquote($target));
+
+    my $logfunc = sub {
+	my $line = shift;
+	log_info($self, "tar: $line");
+    };
+
+    PVE::Tools::run_command($cmd, logfunc => $logfunc);
+
+    return;
+};
+
+# NOTE This only serves as an example to illustrate the 'directory' restore mechanism. It is not
+# fleshed out properly, e.g. I didn't check if exclusion is compatible with
+# proxmox-backup-client/rsync or xattrs/ACL/etc. work as expected!
+my sub backup_directory_squashfs {
+    my ($self, $vmid, $directory, $exclude_patterns, $bandwidth_limit) = @_;
+
+    my $target = "$self->{$vmid}->{'backup-dir'}/archive.sqfs";
+
+    my $mksquashfs = ['mksquashfs', $directory, $target, '-quiet', '-no-progress'];
+
+    push $mksquashfs->@*, '-wildcards';
+
+    for my $pattern ($exclude_patterns->@*) {
+	if ($pattern !~ m|^/|) { # non-anchored
+	    push $mksquashfs->@*, '-e', "... $pattern";
+	} else { # anchored
+	    push $mksquashfs->@*, '-e', substr($pattern, 1); # need to strip leading slash
+	}
+    }
+
+    my $cmd = [ $mksquashfs ];
+
+    push @$cmd, [ 'cstream', '-t', $bandwidth_limit * 1024 ] if $bandwidth_limit;
+
+    my $logfunc = sub {
+	my $line = shift;
+	log_info($self, "mksquashfs: $line");
+    };
+
+    PVE::Tools::run_command($cmd, logfunc => $logfunc);
+
+    return;
+};
+
+sub backup_container {
+    my ($self, $vmid, $guest_config, $exclude_patterns, $info) = @_;
+
+    my $target = "$self->{$vmid}->{'backup-dir'}/guest.conf";
+    file_set_contents($target, $guest_config);
+
+    if (my $firewall_config = $info->{'firewall-config'}) {
+	$target = "$self->{$vmid}->{'backup-dir'}/firewall.conf";
+	file_set_contents($target, $firewall_config);
+    }
+
+    my $backup_mode = $self->{'storage-plugin'}->get_lxc_backup_mode($self->{scfg});
+    if ($backup_mode eq 'tar') {
+	backup_directory_tar(
+	    $self,
+	    $vmid,
+	    $info->{directory},
+	    $exclude_patterns,
+	    $info->{sources},
+	    $info->{'bandwidth-limit'},
+	);
+    } elsif ($backup_mode eq 'squashfs') {
+	backup_directory_squashfs(
+	    $self,
+	    $vmid,
+	    $info->{directory},
+	    $exclude_patterns,
+	    $info->{'bandwidth-limit'},
+	);
+    } else {
+	die "got unexpected backup mode '$backup_mode' from storage plugin\n";
+    }
+}
+
+# Restore API
+
+sub restore_get_mechanism {
+    my ($self, $volname) = @_;
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    my ($vmtype) = $relative_backup_dir =~ m!^\d+/([a-z]+)-!;
+
+    return ('qemu-img', $vmtype) if $vmtype eq 'qemu';
+
+    if ($vmtype eq 'lxc') {
+	my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+
+	if (-e "$self->{scfg}->{path}/${relative_backup_dir}/archive.tar") {
+	    $self->{'restore-mechanisms'}->{$volname} = 'tar';
+	    return ('tar', $vmtype);
+	}
+
+	if (-e "$self->{scfg}->{path}/${relative_backup_dir}/archive.sqfs") {
+	    $self->{'restore-mechanisms'}->{$volname} = 'directory';
+	    return ('directory', $vmtype)
+	}
+
+	die "unable to find archive '$volname'\n";
+    }
+
+    die "cannot restore unexpected guest type '$vmtype'\n";
+}
+
+sub archive_get_guest_config {
+    my ($self, $volname) = @_;
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $filename = "$self->{scfg}->{path}/${relative_backup_dir}/guest.conf";
+
+    return file_get_contents($filename);
+}
+
+sub archive_get_firewall_config {
+    my ($self, $volname) = @_;
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $filename = "$self->{scfg}->{path}/${relative_backup_dir}/firewall.conf";
+
+    return if !-e $filename;
+
+    return file_get_contents($filename);
+}
+
+sub restore_vm_init {
+    my ($self, $volname) = @_;
+
+    my $res = {};
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $backup_dir = "$self->{scfg}->{path}/${relative_backup_dir}";
+
+    my @backup_files = glob("$backup_dir/*");
+    for my $backup_file (@backup_files) {
+	next if $backup_file !~ m!^(.*/(.*)\.qcow2)$!;
+	$backup_file = $1; # untaint
+	$res->{$2}->{size} = PVE::Storage::Plugin::file_size_info($backup_file, undef, 'qcow2');
+    }
+
+    return $res;
+}
+
+sub restore_vm_cleanup {
+    my ($self, $volname) = @_;
+
+    return; # nothing to do
+}
+
+sub restore_vm_volume_init {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $image = "$self->{scfg}->{path}/${relative_backup_dir}/${device_name}.qcow2";
+    # NOTE Backing files are not allowed by Proxmox VE when restoring. The reason is that an
+    # untrusted qcow2 image can specify an arbitrary backing file and thus leak data from the host.
+    # For the sake of the directory example plugin, an NBD export is created, but this side-steps
+    # the check and would allow the attack again. An actual implementation should check that the
+    # backing file (or rather, the whole backing chain) is safe first!
+    my $nbd_node = bind_next_free_dev_nbd_node([$image]);
+    $self->{"${volname}/${device_name}"}->{'nbd-node'} = $nbd_node;
+    return {
+	'qemu-img-path' => $nbd_node,
+    };
+}
+
+sub restore_vm_volume_cleanup {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    if (my $nbd_node = delete($self->{"${volname}/${device_name}"}->{'nbd-node'})) {
+	PVE::Tools::run_command(['qemu-nbd', '-d', $nbd_node]);
+    }
+
+    return;
+}
+
+my sub restore_tar_init {
+    my ($self, $volname) = @_;
+
+    my (undef, $relative_backup_dir) = $self->{'storage-plugin'}->parse_volname($volname);
+    return { 'tar-path' => "$self->{scfg}->{path}/${relative_backup_dir}/archive.tar" };
+}
+
+my sub restore_directory_init {
+    my ($self, $volname) = @_;
+
+    my (undef, $relative_backup_dir, $vmid) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $archive = "$self->{scfg}->{path}/${relative_backup_dir}/archive.sqfs";
+
+    my $mount_point = "/run/backup-provider-example/${vmid}.mount";
+    make_path($mount_point);
+    die "unable to create directory $mount_point\n" if !-d $mount_point;
+
+    run_command(['mount', '-o', 'ro', $archive, $mount_point]);
+
+    return { 'archive-directory' => $mount_point };
+}
+
+my sub restore_directory_cleanup {
+    my ($self, $volname) = @_;
+
+    my (undef, undef, $vmid) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $mount_point = "/run/backup-provider-example/${vmid}.mount";
+
+    run_command(['umount', $mount_point]);
+
+    return;
+}
+
+sub restore_container_init {
+    my ($self, $volname, $info) = @_;
+
+    if ($self->{'restore-mechanisms'}->{$volname} eq 'tar') {
+	return restore_tar_init($self, $volname);
+    } elsif ($self->{'restore-mechanisms'}->{$volname} eq 'directory') {
+	return restore_directory_init($self, $volname);
+    } else {
+	die "no restore mechanism set for '$volname'\n";
+    }
+}
+
+sub restore_container_cleanup {
+    my ($self, $volname, $info) = @_;
+
+    if ($self->{'restore-mechanisms'}->{$volname} eq 'tar') {
+	return; # nothing to do
+    } elsif ($self->{'restore-mechanisms'}->{$volname} eq 'directory') {
+	return restore_directory_cleanup($self, $volname);
+    } else {
+	die "no restore mechanism set for '$volname'\n";
+    }
+}
+
+1;
diff --git a/src/PVE/BackupProvider/Plugin/Makefile b/src/PVE/BackupProvider/Plugin/Makefile
index bbd7431..bedc26e 100644
--- a/src/PVE/BackupProvider/Plugin/Makefile
+++ b/src/PVE/BackupProvider/Plugin/Makefile
@@ -1,4 +1,4 @@
-SOURCES = Base.pm
+SOURCES = Base.pm DirectoryExample.pm
 
 .PHONY: install
 install:
diff --git a/src/PVE/Storage/Custom/BackupProviderDirExamplePlugin.pm b/src/PVE/Storage/Custom/BackupProviderDirExamplePlugin.pm
new file mode 100644
index 0000000..d04d9d1
--- /dev/null
+++ b/src/PVE/Storage/Custom/BackupProviderDirExamplePlugin.pm
@@ -0,0 +1,308 @@
+package PVE::Storage::Custom::BackupProviderDirExamplePlugin;
+
+use strict;
+use warnings;
+
+use File::Basename qw(basename);
+
+use PVE::BackupProvider::Plugin::DirectoryExample;
+use PVE::Tools;
+
+use base qw(PVE::Storage::Plugin);
+
+# Helpers
+
+sub get_vm_backup_mechanism {
+    my ($class, $scfg) = @_;
+
+    return $scfg->{'vm-backup-mechanism'} // properties()->{'vm-backup-mechanism'}->{'default'};
+}
+
+sub get_vm_backup_mode {
+    my ($class, $scfg) = @_;
+
+    return $scfg->{'vm-backup-mode'} // properties()->{'vm-backup-mode'}->{'default'};
+}
+
+sub get_lxc_backup_mode {
+    my ($class, $scfg) = @_;
+
+    return $scfg->{'lxc-backup-mode'} // properties()->{'lxc-backup-mode'}->{'default'};
+}
+
+# Configuration
+
+sub api {
+    return 11;
+}
+
+sub type {
+    return 'backup-provider-dir-example';
+}
+
+sub plugindata {
+    return {
+	content => [ { backup => 1, none => 1 }, { backup => 1 } ],
+	features => { 'backup-provider' => 1 },
+	'sensitive-properties' => {},
+    };
+}
+
+sub properties {
+    return {
+	'lxc-backup-mode' => {
+	    description => "How to create LXC backups. tar - create a tar archive."
+		." squashfs - create a squashfs image. Requires squashfs-tools to be installed.",
+	    type => 'string',
+	    enum => [qw(tar squashfs)],
+	    default => 'tar',
+	},
+	'vm-backup-mechanism' => {
+	    description => "Which mechanism to use for creating VM backups. nbd - access data via "
+		." NBD export. file-handle - access data via file handle.",
+	    type => 'string',
+	    enum => [qw(nbd file-handle)],
+	    default => 'file-handle',
+	},
+	'vm-backup-mode' => {
+	    description => "How to create VM backups. full - always create full backups."
+		." incremental - create incremental backups when possible, fallback to full when"
+		." necessary, e.g. VM disk's bitmap is invalid.",
+	    type => 'string',
+	    enum => [qw(full incremental)],
+	    default => 'full',
+	},
+    };
+}
+
+sub options {
+    return {
+	path => { fixed => 1 },
+	'lxc-backup-mode' => { optional => 1 },
+	'vm-backup-mechanism' => { optional => 1 },
+	'vm-backup-mode' => { optional => 1 },
+	disable => { optional => 1 },
+	nodes => { optional => 1 },
+	'prune-backups' => { optional => 1 },
+	'max-protected-backups' => { optional => 1 },
+    };
+}
+
+# Storage implementation
+
+# NOTE a proper backup storage should implement this
+sub prune_backups {
+    my ($class, $scfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc) = @_;
+
+    die "not implemented";
+}
+
+sub parse_volname {
+    my ($class, $volname) = @_;
+
+    if ($volname =~ m!^backup/((\d+)/[a-z]+-\d+)$!) {
+	my ($filename, $vmid) = ($1, $2);
+	return ('backup', $filename, $vmid);
+    }
+
+    die "unable to parse volume name '$volname'\n";
+}
+
+sub path {
+    my ($class, $scfg, $volname, $storeid, $snapname) = @_;
+
+    die "volume snapshot is not possible on backup-provider-dir-example volume" if $snapname;
+
+    my ($type, $filename, $vmid) = $class->parse_volname($volname);
+
+    return ("$scfg->{path}/${filename}", $vmid, $type);
+}
+
+sub create_base {
+    my ($class, $storeid, $scfg, $volname) = @_;
+
+    die "cannot create base image in backup-provider-dir-example storage\n";
+}
+
+sub clone_image {
+    my ($class, $scfg, $storeid, $volname, $vmid, $snap) = @_;
+
+    die "can't clone images in backup-provider-dir-example storage\n";
+}
+
+sub alloc_image {
+    my ($class, $storeid, $scfg, $vmid, $fmt, $name, $size) = @_;
+
+    die "can't allocate space in backup-provider-dir-example storage\n";
+}
+
+# NOTE a proper backup storage should implement this
+sub free_image {
+    my ($class, $storeid, $scfg, $volname, $isBase) = @_;
+
+    # if it's a backing file, it would need to be merged into the upper image first.
+
+    die "not implemented";
+}
+
+sub list_images {
+    my ($class, $storeid, $scfg, $vmid, $vollist, $cache) = @_;
+
+    my $res = [];
+
+    return $res;
+}
+
+sub list_volumes {
+    my ($class, $storeid, $scfg, $vmid, $content_types) = @_;
+
+    my $path = $scfg->{path};
+
+    my $res = [];
+    for my $type ($content_types->@*) {
+	next if $type ne 'backup';
+
+	my @guest_dirs = glob("$path/*");
+	for my $guest_dir (@guest_dirs) {
+	    next if !-d $guest_dir || $guest_dir !~ m!/(\d+)$!;
+
+	    my $backup_vmid = basename($guest_dir);
+
+	    next if defined($vmid) && $backup_vmid != $vmid;
+
+	    my @backup_dirs = glob("$guest_dir/*");
+	    for my $backup_dir (@backup_dirs) {
+		next if !-d $backup_dir || $backup_dir !~ m!/(lxc|qemu)-(\d+)$!;
+		my ($subtype, $backup_id) = ($1, $2);
+
+		my $size = 0;
+		my @backup_files = glob("$backup_dir/*");
+		$size += -s $_ for @backup_files;
+
+		push $res->@*, {
+		    volid => "$storeid:backup/${backup_vmid}/${subtype}-${backup_id}",
+		    vmid => $backup_vmid,
+		    format => "directory",
+		    ctime => $backup_id,
+		    size => $size,
+		    subtype => $subtype,
+		    content => $type,
+		    # TODO parent for incremental
+		};
+	    }
+	}
+    }
+
+    return $res;
+}
+
+sub activate_storage {
+    my ($class, $storeid, $scfg, $cache) = @_;
+
+    my $path = $scfg->{path};
+
+    my $timeout = 2;
+    if (!PVE::Tools::run_fork_with_timeout($timeout, sub {-d $path})) {
+	die "unable to activate storage '$storeid' - directory '$path' does not exist or is"
+	    ." unreachable\n";
+    }
+
+    return 1;
+}
+
+sub deactivate_storage {
+    my ($class, $storeid, $scfg, $cache) = @_;
+
+    return 1;
+}
+
+sub activate_volume {
+    my ($class, $storeid, $scfg, $volname, $snapname, $cache) = @_;
+
+    die "volume snapshot is not possible on backup-provider-dir-example volume" if $snapname;
+
+    return 1;
+}
+
+sub deactivate_volume {
+    my ($class, $storeid, $scfg, $volname, $snapname, $cache) = @_;
+
+    die "volume snapshot is not possible on backup-provider-dir-example volume" if $snapname;
+
+    return 1;
+}
+
+sub get_volume_attribute {
+    my ($class, $scfg, $storeid, $volname, $attribute) = @_;
+
+    return;
+}
+
+# NOTE a proper backup storage should implement this to support backup notes and
+# setting protected status.
+sub update_volume_attribute {
+    my ($class, $scfg, $storeid, $volname, $attribute, $value) = @_;
+
+    die "attribute '$attribute' is not supported on backup-provider-dir-example volume";
+}
+
+sub volume_size_info {
+    my ($class, $scfg, $storeid, $volname, $timeout) = @_;
+
+    my (undef, $relative_backup_dir) = $class->parse_volname($volname);
+    my ($ctime) = $relative_backup_dir =~ m/-(\d+)$/;
+    my $backup_dir = "$scfg->{path}/${relative_backup_dir}";
+
+    my $size = 0;
+    my @backup_files = glob("$backup_dir/*");
+    for my $backup_file (@backup_files) {
+	if ($backup_file =~ m!\.qcow2$!) {
+	    $size += $class->file_size_info($backup_file, undef, 'qcow2');
+	} else {
+	    $size += -s $backup_file;
+	}
+    }
+
+    my $parent; # TODO for incremental
+
+    return wantarray ? ($size, 'directory', $size, $parent, $ctime) : $size;
+}
+
+sub volume_resize {
+    my ($class, $scfg, $storeid, $volname, $size, $running) = @_;
+
+    die "volume resize is not possible on backup-provider-dir-example volume";
+}
+
+sub volume_snapshot {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot is not possible on backup-provider-dir-example volume";
+}
+
+sub volume_snapshot_rollback {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot rollback is not possible on backup-provider-dir-example volume";
+}
+
+sub volume_snapshot_delete {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot delete is not possible on backup-provider-dir-example volume";
+}
+
+sub volume_has_feature {
+    my ($class, $scfg, $feature, $storeid, $volname, $snapname, $running) = @_;
+
+    return 0;
+}
+
+sub new_backup_provider {
+    my ($class, $scfg, $storeid, $bandwidth_limit, $log_function) = @_;
+
+    return PVE::BackupProvider::Plugin::DirectoryExample->new(
+	$class, $scfg, $storeid, $bandwidth_limit, $log_function);
+}
+
+1;
diff --git a/src/PVE/Storage/Custom/Makefile b/src/PVE/Storage/Custom/Makefile
new file mode 100644
index 0000000..c1e3eca
--- /dev/null
+++ b/src/PVE/Storage/Custom/Makefile
@@ -0,0 +1,5 @@
+SOURCES = BackupProviderDirExamplePlugin.pm
+
+.PHONY: install
+install:
+	for i in ${SOURCES}; do install -D -m 0644 $$i ${DESTDIR}${PERLDIR}/PVE/Storage/Custom/$$i; done
diff --git a/src/PVE/Storage/Makefile b/src/PVE/Storage/Makefile
index ce3fd68..fc0431f 100644
--- a/src/PVE/Storage/Makefile
+++ b/src/PVE/Storage/Makefile
@@ -21,4 +21,5 @@ SOURCES= \
 install:
 	make -C Common install
 	for i in ${SOURCES}; do install -D -m 0644 $$i ${DESTDIR}${PERLDIR}/PVE/Storage/$$i; done
+	make -C Custom install
 	make -C LunCmd install
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [POC storage v6 17/37] Borg example plugin

[Fiona Ebner]

Archive names start with the guest type and ID and then the same
timestamp format as PBS.

Container archives have the following structure:
guest.config
firewall.config
filesystem/ # containing the whole filesystem structure

VM archives have the following structure
guest.config
firewall.config
volumes/ # containing a raw file for each device

A bindmount (respectively symlinks) are used to achieve this
structure, because Borg doesn't seem to support renaming on-the-fly.
(Prefix stripping via the "slashdot hack" would have helped slightly,
but is only in Borg >= 1.4
https://github.com/borgbackup/borg/actions/runs/7967940995)

NOTE: Bandwidth limit is not yet honored and the task size is not
calculated yet. Discard for VM backups would also be nice to have, but
it's not entirely clear how (parsing progress and discarding according
to that is one idea). There is no dirty bitmap support, not sure if
that is feasible to add.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Remove superfluous $storeid parameter from methods. It's already
  part of $self.
* Rename restore_get_*_config() methods to archive_get_*_config().
* Replace job/backup hooks with dedicated methods.
* Remove backup_get_archive_name() method. The archive name is now
  returned as part of the backup_init() method.
* Supersede backup_get_task_size() method and return
  extensible-in-the-future stats as part of the backup_cleanup()
  method.

 src/PVE/BackupProvider/Plugin/Borg.pm      | 460 ++++++++++++++
 src/PVE/BackupProvider/Plugin/Makefile     |   2 +-
 src/PVE/Storage/Custom/BorgBackupPlugin.pm | 689 +++++++++++++++++++++
 src/PVE/Storage/Custom/Makefile            |   3 +-
 4 files changed, 1152 insertions(+), 2 deletions(-)
 create mode 100644 src/PVE/BackupProvider/Plugin/Borg.pm
 create mode 100644 src/PVE/Storage/Custom/BorgBackupPlugin.pm

diff --git a/src/PVE/BackupProvider/Plugin/Borg.pm b/src/PVE/BackupProvider/Plugin/Borg.pm
new file mode 100644
index 0000000..fac8dd9
--- /dev/null
+++ b/src/PVE/BackupProvider/Plugin/Borg.pm
@@ -0,0 +1,460 @@
+package PVE::BackupProvider::Plugin::Borg;
+
+use strict;
+use warnings;
+
+use File::chdir;
+use File::Basename qw(basename);
+use File::Path qw(make_path remove_tree);
+use POSIX qw(strftime);
+
+use PVE::Tools;
+
+# ($vmtype, $vmid, $time_string)
+our $ARCHIVE_RE_3 = qr!^pve-(lxc|qemu)-([0-9]+)-([0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}Z)$!;
+
+sub archive_name {
+    my ($vmtype, $vmid, $backup_time) = @_;
+
+    return "pve-${vmtype}-${vmid}-" . strftime("%FT%TZ", gmtime($backup_time));
+}
+
+# remove_tree can be very verbose by default, do explicit error handling and limit to one message
+my sub _remove_tree {
+    my ($path) = @_;
+
+    remove_tree($path, { error => \my $err });
+    if ($err && @$err) { # empty array if no error
+	for my $diag (@$err) {
+	    my ($file, $message) = %$diag;
+	    die "cannot remove_tree '$path': $message\n" if $file eq '';
+	    die "cannot remove_tree '$path': unlinking $file failed - $message\n";
+	}
+    }
+}
+
+my sub prepare_run_dir {
+    my ($storeid, $archive, $operation, $uid) = @_;
+
+    my $run_dir = "/run/pve-storage/borg-plugin/${storeid}.${archive}.${operation}.$$";
+    _remove_tree($run_dir);
+    make_path($run_dir) or die "unable to create directory $run_dir\n";
+    chmod(0700, $run_dir) or die "unable to chmod directory $run_dir - $!\n";
+    if ($uid) {
+	chown($uid, -1, $run_dir) or die "unable to change owner for $run_dir\n";
+    }
+
+    return $run_dir;
+}
+
+my sub log_info {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('info', $message);
+}
+
+my sub log_warning {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('warn', $message);
+}
+
+my sub log_error {
+    my ($self, $message) = @_;
+
+    $self->{'log-function'}->('err', $message);
+}
+
+my sub file_contents_from_archive {
+    my ($self, $archive, $file) = @_;
+
+    return $self->{'storage-plugin'}->borg_cmd_extract_contents(
+	$self->{scfg},
+	$self->{storeid},
+	$archive,
+	[$file],
+    );
+}
+
+# Plugin implementation
+
+sub new {
+    my ($class, $storage_plugin, $scfg, $storeid, $log_function) = @_;
+
+    my $self = bless {
+	scfg => $scfg,
+	storeid => $storeid,
+	'storage-plugin' => $storage_plugin,
+	'log-function' => $log_function,
+    }, $class;
+
+    return $self;
+}
+
+sub provider_name {
+    my ($self) = @_;
+
+    return "Borg";
+}
+
+sub job_init {
+    my ($self, $start_time) = @_;
+
+    $self->{'job-id'} = $start_time;
+    $self->{password} = $self->{'storage-plugin'}->borg_get_password(
+	$self->{scfg}, $self->{storeid});
+    $self->{'ssh-key-fh'} = $self->{'storage-plugin'}->borg_open_ssh_key(
+	$self->{scfg}, $self->{storeid});
+}
+
+sub job_cleanup {
+    my ($self) = @_;
+
+    delete($self->{password});
+    close($self->{'ssh-key-fh'});
+
+    return;
+}
+
+sub backup_init {
+    my ($self, $vmid, $vmtype, $start_time) = @_;
+
+    $self->{$vmid}->{archive} = archive_name($vmtype, $vmid, $start_time);
+
+    return { 'archive-name' => $self->{$vmid}->{archive} };
+}
+
+sub backup_cleanup {
+    my ($self, $vmid, $vmtype, $success, $info) = @_;
+
+    if (defined($vmtype) && $vmtype eq 'lxc') {
+	if (my $run_dir = $self->{$vmid}->{'run-dir'}) {
+	    eval {
+		# tmpfs for temporary SSH files gets mounted there in backup_container()
+		eval { PVE::Tools::run_command(['umount', "${run_dir}/ssh"]); };
+		eval { PVE::Tools::run_command(['umount', '-R', "${run_dir}/backup/filesystem"]); };
+		_remove_tree($run_dir);
+	    };
+	    die "unable to clean up $run_dir - $@" if $@;
+	}
+    }
+    return { stats => { 'archive-size' => 0 } }; # TODO get size
+}
+
+sub backup_container_prepare {
+    my ($self, $vmid, $info) = @_;
+
+    my $archive = $self->{$vmid}->{archive};
+    my $run_dir = prepare_run_dir(
+	$self->{storeid}, $archive, "backup-container", $info->{'backup-user-id'});
+    $self->{$vmid}->{'run-dir'} = $run_dir;
+
+    my $create_dir = sub {
+	my $dir = shift;
+	make_path($dir) or die "unable to create directory $dir\n";
+	chmod(0700, $dir) or die "unable to chmod directory $dir\n";
+	chown($info->{'backup-user-id'}, -1, $dir)
+	    or die "unable to change owner for $dir\n";
+    };
+
+    $create_dir->("${run_dir}/backup/");
+    $create_dir->("${run_dir}/backup/filesystem");
+    $create_dir->("${run_dir}/ssh");
+    $create_dir->("${run_dir}/.config");
+    $create_dir->("${run_dir}/.cache");
+
+    for my $subdir ($info->{sources}->@*) {
+	PVE::Tools::run_command([
+	    'mount',
+	    '-o', 'bind,ro',
+	    "$info->{directory}/${subdir}",
+	    "${run_dir}/backup/filesystem/${subdir}",
+	]);
+    }
+}
+
+sub backup_get_mechanism {
+    my ($self, $vmid, $vmtype) = @_;
+
+    return ('file-handle', undef) if $vmtype eq 'qemu';
+    return ('directory', undef) if $vmtype eq 'lxc';
+
+    die "unsupported VM type '$vmtype'\n";
+}
+
+sub backup_handle_log_file {
+    my ($self, $vmid, $filename) = @_;
+
+    return; # don't upload, Proxmox VE keeps the task log too
+}
+
+my sub backup_vm_setup_loopdev {
+    my ($file) = @_;
+
+    my $device;
+    my $parser = sub {
+	my $line = shift;
+	if ($line =~ m@^(/dev/loop\d+)$@) {
+	    $device = $1;
+	}
+    };
+    my $losetup_cmd = [
+	'losetup',
+	'--show',
+	'-r',
+	'-f',
+	$file,
+    ];
+    PVE::Tools::run_command($losetup_cmd, outfunc => $parser);
+    return $device;
+}
+
+sub backup_vm {
+    my ($self, $vmid, $guest_config, $volumes, $info) = @_;
+
+    # TODO honor bandwith limit
+    # TODO discard?
+
+    my $archive = $self->{$vmid}->{archive};
+
+    my $run_dir = prepare_run_dir($self->{storeid}, $archive, "backup-vm");
+    my $volume_dir = "${run_dir}/volumes";
+    make_path($volume_dir) or die "unable to create directory $volume_dir\n";
+
+    PVE::Tools::file_set_contents("${run_dir}/guest.config", $guest_config);
+    my $paths = ['./guest.config'];
+
+    if (my $firewall_config = $info->{'firewall-config'}) {
+	PVE::Tools::file_set_contents("${run_dir}/firewall.config", $firewall_config);
+	push $paths->@*, './firewall.config';
+    }
+
+    my @blockdevs = ();
+
+    # TODO --stats for size?
+
+    eval {
+	for my $device_name (sort keys $volumes->%*) {
+	    # FIXME there is no option to follow symlinks except in combination with special files,
+	    # so loop devices are set up here for this purpose. Newer versions of Borg (since 1.4)
+	    # could use the slashdot hack instead:
+	    # https://github.com/borgbackup/borg/commit/e7bd18d7f38ddf9e58a4587ae4a2ad8a24d67374
+	    my $path = "/proc/$$/fd/" . fileno($volumes->{$device_name}->{'file-handle'});
+	    my $blockdev = backup_vm_setup_loopdev($path);
+	    push @blockdevs, $blockdev;
+
+	    my $link_name = "${volume_dir}/${device_name}.raw";
+	    symlink($blockdev, $link_name) or die "could not create symlink $link_name -> $blockdev\n";
+	    push $paths->@*, "./volumes/" . basename($link_name, ());
+	}
+
+	local $CWD = $run_dir;
+
+	$self->{'storage-plugin'}->borg_cmd_create(
+	    $self->{scfg},
+	    $self->{storeid},
+	    $self->{$vmid}->{archive},
+	    $paths,
+	    ['--read-special', '--progress'],
+	);
+    };
+    my $err = $@;
+    for my $blockdev (@blockdevs) {
+	eval { PVE::Tools::run_command(['losetup', '-d', $blockdev]); };
+	log_warning($self, "cannot cleanup loop device - $@") if $@;
+    }
+    eval { _remove_tree($run_dir) };
+    log_warning($self, $@) if $@;
+    die $err if $err;
+}
+
+sub backup_container {
+    my ($self, $vmid, $guest_config, $exclude_patterns, $info) = @_;
+
+    # TODO honor bandwith limit
+
+    my $run_dir = $self->{$vmid}->{'run-dir'};
+    my $backup_dir = "${run_dir}/backup";
+
+    my $archive = $self->{$vmid}->{archive};
+
+    my $ssh_key;
+    if ($self->{'ssh-key-fh'}) {
+	$ssh_key =
+	    PVE::Tools::safe_read_from($self->{'ssh-key-fh'}, 1024 * 1024, 0, "SSH key file");
+    }
+
+    my (undef, $ssh_options) =
+	$self->{'storage-plugin'}->borg_setup_ssh_dir($self->{scfg}, "${run_dir}/ssh", $ssh_key);
+
+    PVE::Tools::file_set_contents("${backup_dir}/guest.config", $guest_config);
+    my $paths = ['./guest.config'];
+
+    if (my $firewall_config = $info->{'firewall-config'}) {
+	PVE::Tools::file_set_contents("${backup_dir}/firewall.config", $firewall_config);
+	push $paths->@*, './firewall.config';
+    }
+
+    push $paths->@*, "./filesystem";
+
+    my $opts = ['--numeric-ids', '--sparse', '--progress'];
+
+    for my $pattern ($exclude_patterns->@*) {
+	if ($pattern =~ m|^/|) {
+	    push $opts->@*, '-e', "filesystem${pattern}";
+	} else {
+	    push $opts->@*, '-e', "filesystem/**${pattern}";
+	}
+    }
+
+    push $opts->@*, '-e', "filesystem/**lost+found" if $info->{'backup-user-id'} != 0;
+
+    # TODO --stats for size?
+
+    # Don't make it local to avoid permission denied error when changing back, because the method is
+    # executed in a user namespace.
+    $CWD = $backup_dir if $info->{'backup-user-id'} != 0;
+    {
+	local $CWD = $backup_dir;
+	local $ENV{BORG_BASE_DIR} = ${run_dir};
+	local $ENV{BORG_PASSPHRASE} = $self->{password};
+
+	local $ENV{BORG_RSH} = "ssh " . join(" ", $ssh_options->@*);
+
+	my $uri = $self->{'storage-plugin'}->borg_repository_uri($self->{scfg}, $self->{storeid});
+	my $archive = $self->{$vmid}->{archive};
+
+	my $cmd = ['borg', 'create', $opts->@*, "${uri}::${archive}", $paths->@*];
+
+	PVE::Tools::run_command($cmd, errmsg => "command @$cmd failed");
+    }
+}
+
+sub restore_get_mechanism {
+    my ($self, $volname) = @_;
+
+    my (undef, $archive) = $self->{'storage-plugin'}->parse_volname($volname);
+    my ($vmtype) = $archive =~ m!^pve-([^\s-]+)!
+	or die "cannot parse guest type from archive name '$archive'\n";
+
+    return ('qemu-img', $vmtype) if $vmtype eq 'qemu';
+    return ('directory', $vmtype) if $vmtype eq 'lxc';
+
+    die "unexpected guest type '$vmtype'\n";
+}
+
+sub archive_get_guest_config {
+    my ($self, $volname) = @_;
+
+    my (undef, $archive) = $self->{'storage-plugin'}->parse_volname($volname);
+    return file_contents_from_archive($self, $archive, 'guest.config');
+}
+
+sub archive_get_firewall_config {
+    my ($self, $volname) = @_;
+
+    my (undef, $archive) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $config = eval {
+	file_contents_from_archive($self, $archive, 'firewall.config');
+    };
+    if (my $err = $@) {
+	return if $err =~ m!Include pattern 'firewall\.config' never matched\.!;
+	die $err;
+    }
+    return $config;
+}
+
+sub restore_vm_init {
+    my ($self, $volname) = @_;
+
+    my $res = {};
+
+    my (undef, $archive, $vmid) = $self->{'storage-plugin'}->parse_volname($volname);
+
+    my $run_dir = prepare_run_dir($self->{storeid}, $archive, "restore-vm");
+    $self->{$volname}->{'run-dir'} = $run_dir;
+
+    my $mount_point = "${run_dir}/mount";
+    make_path($mount_point) or die "unable to create directory $mount_point\n";
+    $self->{$volname}->{'mount-point'} = $mount_point;
+
+    $self->{'storage-plugin'}->borg_cmd_mount(
+	$self->{scfg},
+	$self->{storeid},
+	$archive,
+	$mount_point,
+    );
+
+    my @backup_files = glob("$mount_point/volumes/*");
+    for my $backup_file (@backup_files) {
+	next if $backup_file !~ m!^(.*/(.*)\.raw)$!; # untaint
+	($backup_file, my $device_name) = ($1, $2);
+	# TODO avoid dependency on base plugin?
+	$res->{$device_name}->{size} =
+	    PVE::Storage::Plugin::file_size_info($backup_file, undef, 'raw');
+    }
+
+    return $res;
+}
+
+sub restore_vm_cleanup {
+    my ($self, $volname) = @_;
+
+    my $run_dir = $self->{$volname}->{'run-dir'} or return;
+    my $mount_point = $self->{$volname}->{'mount-point'};
+
+    eval { PVE::Tools::run_command(['umount', $mount_point]) };
+    eval { _remove_tree($run_dir); };
+    die "unable to clean up $run_dir - $@" if $@;
+
+    return;
+}
+
+sub restore_vm_volume_init {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    my $mount_point = $self->{$volname}->{'mount-point'}
+	or die "expected mount point for archive not present\n";
+
+    return { 'qemu-img-path' => "${mount_point}/volumes/${device_name}.raw" };
+}
+
+sub restore_vm_volume_cleanup {
+    my ($self, $volname, $device_name, $info) = @_;
+
+    return;
+}
+
+sub restore_container_init {
+    my ($self, $volname, $info) = @_;
+
+    my (undef, $archive, $vmid) = $self->{'storage-plugin'}->parse_volname($volname);
+    my $run_dir = prepare_run_dir($self->{storeid}, $archive, "restore-container");
+    $self->{$volname}->{'run-dir'} = $run_dir;
+
+    my $mount_point = "${run_dir}/mount";
+    make_path($mount_point) or die "unable to create directory $mount_point\n";
+    $self->{$volname}->{'mount-point'} = $mount_point;
+
+    $self->{'storage-plugin'}->borg_cmd_mount(
+	$self->{scfg},
+	$self->{storeid},
+	$archive,
+	$mount_point,
+    );
+
+    return { 'archive-directory' => "${mount_point}/filesystem" };
+}
+
+sub restore_container_cleanup {
+    my ($self, $volname, $info) = @_;
+
+    my $run_dir = $self->{$volname}->{'run-dir'} or return;
+    my $mount_point = $self->{$volname}->{'mount-point'};
+
+    eval { PVE::Tools::run_command(['umount', $mount_point]) };
+    eval { _remove_tree($run_dir); };
+    die "unable to clean up $run_dir - $@" if $@;
+}
+
+1;
diff --git a/src/PVE/BackupProvider/Plugin/Makefile b/src/PVE/BackupProvider/Plugin/Makefile
index bedc26e..db08c2d 100644
--- a/src/PVE/BackupProvider/Plugin/Makefile
+++ b/src/PVE/BackupProvider/Plugin/Makefile
@@ -1,4 +1,4 @@
-SOURCES = Base.pm DirectoryExample.pm
+SOURCES = Base.pm Borg.pm DirectoryExample.pm
 
 .PHONY: install
 install:
diff --git a/src/PVE/Storage/Custom/BorgBackupPlugin.pm b/src/PVE/Storage/Custom/BorgBackupPlugin.pm
new file mode 100644
index 0000000..84b12b9
--- /dev/null
+++ b/src/PVE/Storage/Custom/BorgBackupPlugin.pm
@@ -0,0 +1,689 @@
+package PVE::Storage::Custom::BorgBackupPlugin;
+
+use strict;
+use warnings;
+
+use Fcntl qw(F_GETFD F_SETFD FD_CLOEXEC);
+use File::Path qw(make_path remove_tree);
+use JSON qw(from_json);
+use MIME::Base64 qw(decode_base64 encode_base64);
+use Net::IP;
+use POSIX;
+
+use PVE::BackupProvider::Plugin::Borg;
+use PVE::Tools;
+
+use base qw(PVE::Storage::Plugin);
+
+sub api {
+    return 11;
+}
+
+sub check_config {
+    my ($class, $sectionId, $config, $create, $skipSchemaCheck) = @_;
+
+    if (my $ssh_public_keys = $config->{'ssh-server-public-keys'}) {
+	if ($ssh_public_keys !~ m/^[A-Za-z0-9+\/]+={0,2}$/) {
+	    $config->{'ssh-server-public-keys'} = encode_base64($ssh_public_keys, '');
+	}
+    }
+
+    return $class->SUPER::check_config($sectionId, $config, $create, $skipSchemaCheck);
+}
+
+sub borg_repository_uri {
+    my ($class, $scfg, $storeid) = @_;
+
+    my $uri = '';
+    my $server = $scfg->{server} or die "no server configured for $storeid\n";
+    my $username = $scfg->{username} or die "no username configured for $storeid\n";
+    my $prefix = "ssh://$username@";
+    $server = "[$server]" if Net::IP::ip_is_ipv6($server);
+    if (my $port = $scfg->{port}) {
+	$uri = "${prefix}${server}:${port}";
+    } else {
+	$uri = "${prefix}${server}";
+    }
+    $uri .= $scfg->{'repository-path'};
+
+    return $uri;
+}
+
+my sub borg_password_file_name {
+    my ($scfg, $storeid) = @_;
+
+    return "/etc/pve/priv/storage/${storeid}.pw";
+}
+
+my sub borg_set_password {
+    my ($scfg, $storeid, $password) = @_;
+
+    my $pwfile = borg_password_file_name($scfg, $storeid);
+    mkdir "/etc/pve/priv/storage";
+
+    PVE::Tools::file_set_contents($pwfile, "$password\n");
+}
+
+my sub borg_delete_password {
+    my ($scfg, $storeid) = @_;
+
+    my $pwfile = borg_password_file_name($scfg, $storeid);
+
+    unlink $pwfile;
+}
+
+sub borg_get_password {
+    my ($class, $scfg, $storeid) = @_;
+
+    my $pwfile = borg_password_file_name($scfg, $storeid);
+
+    return PVE::Tools::file_read_firstline($pwfile);
+}
+
+sub borg_setup_ssh_dir {
+    my ($class, $scfg, $ssh_dir, $ssh_key) = @_;
+
+    my $dir_created;
+    my $ssh_opts = [];
+
+    my $ensure_dir_created = sub {
+	return if $dir_created;
+	# for container backup, it needs to be created while privileged and already exists
+	if (!-d $ssh_dir) {
+	    make_path($ssh_dir) or die "unable to create directory $ssh_dir\n";
+	    chmod(0700, $ssh_dir) or die "unable to chmod directory $ssh_dir\n";
+	}
+	PVE::Tools::run_command(
+	    ['mount', '-t', 'tmpfs', '-o', 'size=1M,mode=0700', 'tmpfs', $ssh_dir]);
+	$dir_created = 1;
+    };
+
+    if ($ssh_key) {
+	$ensure_dir_created->();
+	PVE::Tools::file_set_contents("${ssh_dir}/ssh.key", $ssh_key, 0600);
+	push $ssh_opts->@*, '-i', "${ssh_dir}/ssh.key";
+    }
+
+    if ($scfg->{'ssh-server-public-keys'}) {
+	$ensure_dir_created->();
+	my $raw = decode_base64($scfg->{'ssh-server-public-keys'});
+	PVE::Tools::file_set_contents("${ssh_dir}/known_hosts", $raw, 0600);
+	push $ssh_opts->@*, '-o', "UserKnownHostsFile=${ssh_dir}/known_hosts";
+	push $ssh_opts->@*, '-o', "GlobalKnownHostsFile=none";
+    }
+
+    return ($dir_created, $ssh_opts);
+}
+
+sub borg_cmd_env {
+    my ($class, $scfg, $storeid, $sub) = @_;
+
+    my $ssh_dir = "/run/pve-storage/borg-plugin/${storeid}.ssh.$$";
+    my $ssh_key = borg_get_ssh_key($scfg, $storeid);
+    my ($uses_ssh_dir, $ssh_options) = $class->borg_setup_ssh_dir($scfg, $ssh_dir, $ssh_key);
+    local $ENV{BORG_RSH} = "ssh " . join(" ", $ssh_options->@*);
+
+    local $ENV{BORG_PASSPHRASE} = $class->borg_get_password($scfg, $storeid);
+
+    my $res = eval {
+	my $uri = $class->borg_repository_uri($scfg, $storeid);
+	return $sub->($uri);
+    };
+    my $err = $@;
+
+    if ($uses_ssh_dir) {
+	eval { PVE::Tools::run_command(['umount', "$ssh_dir"]); };
+	warn "unable to unmount directory $ssh_dir - $@" if $@;
+	eval { remove_tree($ssh_dir); };
+	warn "unable to cleanup directory $ssh_dir - $@" if $@;
+    }
+
+    die $err if $err;
+
+    return $res;
+}
+
+sub borg_cmd_list {
+    my ($class, $scfg, $storeid) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $json = '';
+	my $cmd = ['borg', 'list', '--json', $uri];
+
+	my $errfunc = sub { warn $_[0]; };
+	my $outfunc = sub { $json .= $_[0]; };
+
+	PVE::Tools::run_command(
+	    $cmd, errmsg => "command @$cmd failed", outfunc => $outfunc, errfunc => $errfunc);
+
+	my $res = eval { from_json($json) };
+	die "unable to parse 'borg list' output - $@\n" if $@;
+	return $res;
+    });
+}
+
+sub borg_cmd_create {
+    my ($class, $scfg, $storeid, $archive, $paths, $opts) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $cmd = ['borg', 'create', $opts->@*, "${uri}::${archive}", $paths->@*];
+
+	PVE::Tools::run_command($cmd, errmsg => "command @$cmd failed");
+
+	return;
+    });
+}
+
+sub borg_cmd_extract_contents {
+    my ($class, $scfg, $storeid, $archive, $paths) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $output = '';
+	my $outfunc = sub {
+	    $output .= "$_[0]\n";
+	};
+
+	my $cmd = ['borg', 'extract', '--stdout', "${uri}::${archive}", $paths->@*];
+
+	PVE::Tools::run_command($cmd, errmsg => "command @$cmd failed", outfunc => $outfunc);
+
+	return $output;
+    });
+}
+
+sub borg_cmd_delete {
+    my ($class, $scfg, $storeid, $archive) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $cmd = ['borg', 'delete', "${uri}::${archive}"];
+
+	PVE::Tools::run_command($cmd, errmsg => "command @$cmd failed");
+
+	return;
+    });
+}
+
+sub borg_cmd_info {
+    my ($class, $scfg, $storeid, $archive, $timeout) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $json = '';
+	my $cmd = ['borg', 'info', '--json', "${uri}::${archive}"];
+
+	my $errfunc = sub { warn $_[0]; };
+	my $outfunc = sub { $json .= $_[0]; };
+
+	PVE::Tools::run_command(
+	    $cmd,
+	    errmsg => "command @$cmd failed",
+	    timeout => $timeout,
+	    outfunc => $outfunc,
+	    errfunc => $errfunc,
+	);
+
+	my $res = eval { from_json($json) };
+	die "unable to parse 'borg info' output for archive '$archive' - $@\n" if $@;
+	return $res;
+    });
+}
+
+sub borg_cmd_mount {
+    my ($class, $scfg, $storeid, $archive, $mount_point) = @_;
+
+    return $class->borg_cmd_env($scfg, $storeid, sub {
+	my ($uri) = @_;
+
+	my $cmd = ['borg', 'mount', "${uri}::${archive}", $mount_point];
+
+	PVE::Tools::run_command($cmd, errmsg => "command @$cmd failed");
+
+	return;
+    });
+}
+
+my sub parse_backup_time {
+    my ($time_string) = @_;
+
+    my @tm = (POSIX::strptime($time_string, "%FT%TZ"));
+    # expect sec, min, hour, mday, mon, year
+    if (grep { !defined($_) } @tm[0..5]) {
+	warn "error parsing time from string '$time_string'\n";
+	return 0;
+    } else {
+	local $ENV{TZ} = 'UTC'; # time string is UTC
+
+	# Fill in isdst to avoid undef warning. No daylight saving time for UTC.
+	$tm[8] //= 0;
+
+	if (my $since_epoch = mktime(@tm)) {
+	    return int($since_epoch);
+	} else {
+	    warn "error parsing time from string '$time_string'\n";
+	    return 0;
+	}
+    }
+}
+
+# Helpers
+
+sub type {
+    return 'borg';
+}
+
+sub plugindata {
+    return {
+	content => [ { backup => 1, none => 1 }, { backup => 1 } ],
+	features => { 'backup-provider' => 1 },
+	'sensitive-properties' => {
+	    password => 1,
+	    'ssh-private-key' => 1,
+	},
+    };
+}
+
+sub properties {
+    return {
+	'repository-path' => {
+	    description => "Path to the backup repository",
+	    type => 'string',
+	},
+	'ssh-private-key' => {
+	    # Since 1 is written to the config when the key is present, the format is not checked.
+	    description => "SSH identity/private key for the client-side in PEM format.",
+	    type => 'string',
+	},
+	'ssh-server-public-keys' => {
+	    description => "SSH public key(s) for the server-side, (one key per line, OpenSSH"
+		." format).",
+	    type => 'string',
+	},
+    };
+}
+
+sub options {
+    return {
+	'repository-path' => { fixed => 1 },
+	server => { fixed => 1 },
+	port => { optional => 1 },
+	username => { fixed => 1 },
+	'ssh-private-key' => { optional => 1 },
+	'ssh-server-public-keys' => { optional => 1 },
+	password => { optional => 1 },
+	disable => { optional => 1 },
+	nodes => { optional => 1 },
+	'prune-backups' => { optional => 1 },
+	'max-protected-backups' => { optional => 1 },
+    };
+}
+
+sub borg_ssh_key_file_name {
+    my ($scfg, $storeid) = @_;
+
+    return "/etc/pve/priv/storage/${storeid}.ssh.key";
+}
+
+sub borg_set_ssh_key {
+    my ($scfg, $storeid, $key) = @_;
+
+    my $keyfile = borg_ssh_key_file_name($scfg, $storeid);
+    mkdir "/etc/pve/priv/storage";
+
+    PVE::Tools::file_set_contents($keyfile, "$key\n");
+}
+
+sub borg_delete_ssh_key {
+    my ($scfg, $storeid) = @_;
+
+    my $keyfile = borg_ssh_key_file_name($scfg, $storeid);
+
+    if (!unlink $keyfile) {
+	return if $! == ENOENT;
+	die "failed to delete SSH key! $!\n";
+    }
+    delete $scfg->{'ssh-private-key'};
+}
+
+sub borg_get_ssh_key {
+    my ($scfg, $storeid) = @_;
+
+    my $keyfile = borg_ssh_key_file_name($scfg, $storeid);
+
+    return if !-f $keyfile;
+
+    return PVE::Tools::file_get_contents($keyfile);
+}
+
+# Returns a file handle with FD_CLOEXEC disabled if there is an SSH key , or `undef` if there is
+# not. Dies on error.
+sub borg_open_ssh_key {
+    my ($self, $scfg, $storeid) = @_;
+
+    my $ssh_key_file = borg_ssh_key_file_name($scfg, $storeid);
+
+    my $keyfd;
+    if (!open($keyfd, '<', $ssh_key_file)) {
+	if ($! == ENOENT) {
+	    die "SSH key configured but no key file found!\n" if $scfg->{'ssh-private-key'};
+	    return undef;
+	}
+	die "failed to open SSH key: $ssh_key_file: $!\n";
+    }
+    my $flags = fcntl($keyfd, F_GETFD, 0)
+	// die "failed to get file descriptor flags for SSH key FD: $!\n";
+    fcntl($keyfd, F_SETFD, $flags & ~FD_CLOEXEC)
+	or die "failed to remove FD_CLOEXEC from SSH key file descriptor\n";
+
+    return $keyfd;
+}
+
+# Storage implementation
+
+sub on_add_hook {
+    my ($class, $storeid, $scfg, %param) = @_;
+
+    if (defined(my $password = $param{password})) {
+	borg_set_password($scfg, $storeid, $password);
+    } else {
+	borg_delete_password($scfg, $storeid);
+    }
+
+    if (defined(my $ssh_key = delete $param{'ssh-private-key'})) {
+	borg_set_ssh_key($scfg, $storeid, $ssh_key);
+	$scfg->{'ssh-private-key'} = 1;
+    } else {
+	borg_delete_ssh_key($scfg, $storeid);
+    }
+
+    if ($scfg->{'ssh-server-public-keys'}) {
+	my $ssh_public_keys = decode_base64($scfg->{'ssh-server-public-keys'});
+	PVE::Tools::validate_ssh_public_keys($ssh_public_keys);
+    }
+
+    return;
+}
+
+sub on_update_hook {
+    my ($class, $storeid, $scfg, %param) = @_;
+
+    if (exists($param{password})) {
+	if (defined($param{password})) {
+	    borg_set_password($scfg, $storeid, $param{password});
+	} else {
+	    borg_delete_password($scfg, $storeid);
+	}
+    }
+
+    if (exists($param{'ssh-private-key'})) {
+	if (defined(my $ssh_key = delete($param{'ssh-private-key'}))) {
+	    borg_set_ssh_key($scfg, $storeid, $ssh_key);
+	    $scfg->{'ssh-private-key'} = 1;
+	} else {
+	    borg_delete_ssh_key($scfg, $storeid);
+	}
+    }
+
+    if ($scfg->{'ssh-server-public-keys'}) {
+	my $ssh_public_keys = decode_base64($scfg->{'ssh-server-public-keys'});
+	PVE::Tools::validate_ssh_public_keys($ssh_public_keys);
+    }
+
+    return;
+}
+
+sub on_delete_hook {
+    my ($class, $storeid, $scfg) = @_;
+
+    borg_delete_password($scfg, $storeid);
+    borg_delete_ssh_key($scfg, $storeid);
+
+    return;
+}
+
+sub prune_backups {
+    my ($class, $scfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc) = @_;
+
+    # FIXME - is 'borg prune' compatible with ours?
+    die "not implemented";
+}
+
+sub parse_volname {
+    my ($class, $volname) = @_;
+
+    if ($volname =~ m!^backup/(.*)$!) {
+	my $archive = $1;
+	if ($archive =~ $PVE::BackupProvider::Plugin::Borg::ARCHIVE_RE_3) {
+	    return ('backup', $archive, $2);
+	}
+    }
+
+    die "unable to parse Borg volume name '$volname'\n";
+}
+
+sub path {
+    my ($class, $scfg, $volname, $storeid, $snapname) = @_;
+
+    die "volume snapshot is not possible on Borg volume" if $snapname;
+
+    my $uri = $class->borg_repository_uri($scfg, $storeid);
+    my (undef, $archive) = $class->parse_volname($volname);
+
+    return "${uri}::${archive}";
+}
+
+sub create_base {
+    my ($class, $storeid, $scfg, $volname) = @_;
+
+    die "cannot create base image in Borg storage\n";
+}
+
+sub clone_image {
+    my ($class, $scfg, $storeid, $volname, $vmid, $snap) = @_;
+
+    die "can't clone images in Borg storage\n";
+}
+
+sub alloc_image {
+    my ($class, $storeid, $scfg, $vmid, $fmt, $name, $size) = @_;
+
+    die "can't allocate space in Borg storage\n";
+}
+
+sub free_image {
+    my ($class, $storeid, $scfg, $volname, $isBase) = @_;
+
+    my (undef, $archive) = $class->parse_volname($volname);
+
+    borg_cmd_delete($class, $scfg, $storeid, $archive);
+
+    return;
+}
+
+sub list_images {
+    my ($class, $storeid, $scfg, $vmid, $vollist, $cache) = @_;
+
+    return []; # guest images are not supported, only backups
+}
+
+sub list_volumes {
+    my ($class, $storeid, $scfg, $vmid, $content_types) = @_;
+
+    my $res = [];
+
+    return $res if !grep { $_ eq 'backup' } $content_types->@*;
+
+    my $archives = $class->borg_cmd_list($scfg, $storeid)->{archives}
+	or die "expected 'archives' key in 'borg list' JSON output missing\n";
+
+    for my $info ($archives->@*) {
+	my $archive = $info->{archive};
+	my ($vmtype, $backup_vmid, $time_string) =
+	    $archive =~ $PVE::BackupProvider::Plugin::Borg::ARCHIVE_RE_3 or next;
+
+	next if defined($vmid) && $vmid != $backup_vmid;
+
+	push $res->@*, {
+	    volid => "${storeid}:backup/${archive}",
+	    size => 0, # FIXME how to cheaply get?
+	    content => 'backup',
+	    ctime => parse_backup_time($time_string),
+	    vmid => $backup_vmid,
+	    format => "borg-archive",
+	    subtype => $vmtype,
+	}
+    }
+
+    return $res;
+}
+
+sub status {
+    my ($class, $storeid, $scfg, $cache) = @_;
+
+    my $uri = $class->borg_repository_uri($scfg, $storeid);
+
+    my $res;
+
+    if ($uri =~ m!^ssh://!) {
+	#FIXME ssh and df on target?
+	return;
+    } else { # $uri is a local path
+	my $timeout = 2;
+	$res = PVE::Tools::df($uri, $timeout);
+
+	return if !$res || !$res->{total};
+    }
+
+
+    return ($res->{total}, $res->{avail}, $res->{used}, 1);
+}
+
+sub activate_storage {
+    my ($class, $storeid, $scfg, $cache) = @_;
+
+    # TODO how to cheaply check? split ssh and non-ssh?
+
+    return 1;
+}
+
+sub deactivate_storage {
+    my ($class, $storeid, $scfg, $cache) = @_;
+
+    return 1;
+}
+
+sub activate_volume {
+    my ($class, $storeid, $scfg, $volname, $snapname, $cache) = @_;
+
+    die "volume snapshot is not possible on Borg volume" if $snapname;
+
+    return 1;
+}
+
+sub deactivate_volume {
+    my ($class, $storeid, $scfg, $volname, $snapname, $cache) = @_;
+
+    die "volume snapshot is not possible on Borg volume" if $snapname;
+
+    return 1;
+}
+
+sub get_volume_attribute {
+    my ($class, $scfg, $storeid, $volname, $attribute) = @_;
+
+    return;
+}
+
+sub update_volume_attribute {
+    my ($class, $scfg, $storeid, $volname, $attribute, $value) = @_;
+
+    # FIXME notes or protected possible?
+
+    die "attribute '$attribute' is not supported on Borg volume";
+}
+
+sub volume_size_info {
+    my ($class, $scfg, $storeid, $volname, $timeout) = @_;
+
+    my (undef, $archive) = $class->parse_volname($volname);
+    my (undef, undef, $time_string) =
+	$archive =~ $PVE::BackupProvider::Plugin::Borg::ARCHIVE_RE_3;
+
+    my $backup_time = 0;
+    if ($time_string) {
+	$backup_time = parse_backup_time($time_string)
+    } else {
+	warn "could not parse time from archive name '$archive'\n";
+    }
+
+    my $archives = borg_cmd_info($class, $scfg, $storeid, $archive, $timeout)->{archives}
+	or die "expected 'archives' key in 'borg info' JSON output missing\n";
+
+    my $stats = eval { $archives->[0]->{stats} }
+	or die "expected entry in 'borg info' JSON output missing\n";
+    my ($size, $used) = $stats->@{qw(original_size deduplicated_size)};
+
+    ($size) = ($size =~ /^(\d+)$/); # untaint
+    die "size '$size' not an integer\n" if !defined($size);
+    # coerce back from string
+    $size = int($size);
+    ($used) = ($used =~ /^(\d+)$/); # untaint
+    die "used '$used' not an integer\n" if !defined($used);
+    # coerce back from string
+    $used = int($used);
+
+    return wantarray ? ($size, 'borg-archive', $used, undef, $backup_time) : $size;
+}
+
+sub volume_resize {
+    my ($class, $scfg, $storeid, $volname, $size, $running) = @_;
+
+    die "volume resize is not possible on Borg volume";
+}
+
+sub volume_snapshot {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot is not possible on Borg volume";
+}
+
+sub volume_snapshot_rollback {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot rollback is not possible on Borg volume";
+}
+
+sub volume_snapshot_delete {
+    my ($class, $scfg, $storeid, $volname, $snap) = @_;
+
+    die "volume snapshot delete is not possible on Borg volume";
+}
+
+sub volume_has_feature {
+    my ($class, $scfg, $feature, $storeid, $volname, $snapname, $running) = @_;
+
+    return 0;
+}
+
+sub rename_volume {
+    my ($class, $scfg, $storeid, $source_volname, $target_vmid, $target_volname) = @_;
+
+    die "volume rename is not implemented in Borg storage plugin\n";
+}
+
+sub new_backup_provider {
+    my ($class, $scfg, $storeid, $bandwidth_limit, $log_function) = @_;
+
+    return PVE::BackupProvider::Plugin::Borg->new(
+	$class, $scfg, $storeid, $bandwidth_limit, $log_function);
+}
+
+1;
diff --git a/src/PVE/Storage/Custom/Makefile b/src/PVE/Storage/Custom/Makefile
index c1e3eca..886442d 100644
--- a/src/PVE/Storage/Custom/Makefile
+++ b/src/PVE/Storage/Custom/Makefile
@@ -1,4 +1,5 @@
-SOURCES = BackupProviderDirExamplePlugin.pm
+SOURCES = BackupProviderDirExamplePlugin.pm \
+          BorgBackupPlugin.pm
 
 .PHONY: install
 install:
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 18/37] backup: keep track of block-node size for fleecing

[Fiona Ebner]

For fleecing, the size needs to match exactly what QEMU sees. In
particular, EFI disks might be attached with a 'size=' option, meaning
that size can be different from the volume's size. Commit 36377acf
("backup: disk info: also keep track of size") introduced size
tracking and it was used for fleecing since then, but the accurate
size information needs to be queried via QMP.

Should also help with the following issue reported in the community
forum:
https://forum.proxmox.com/threads/152202

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 31 +++++++++++++++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 3238e344..4a258895 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -558,7 +558,7 @@ my sub allocate_fleecing_images {
 		my $name = "vm-$vmid-fleece-$n";
 		$name .= ".$format" if $scfg->{path};
 
-		my $size = PVE::Tools::convert_size($di->{size}, 'b' => 'kb');
+		my $size = PVE::Tools::convert_size($di->{'block-node-size'}, 'b' => 'kb');
 
 		$di->{'fleece-volid'} = PVE::Storage::vdisk_alloc(
 		    $self->{storecfg}, $fleecing_storeid, $vmid, $format, $name, $size);
@@ -607,7 +607,7 @@ my sub attach_fleecing_images {
 	    my $drive = "file=$path,if=none,id=$devid,format=$format,discard=unmap";
 	    # Specify size explicitly, to make it work if storage backend rounded up size for
 	    # fleecing image when allocating.
-	    $drive .= ",size=$di->{size}" if $format eq 'raw';
+	    $drive .= ",size=$di->{'block-node-size'}" if $format eq 'raw';
 	    $drive =~ s/\\/\\\\/g;
 	    my $ret = PVE::QemuServer::Monitor::hmp_cmd($vmid, "drive_add auto \"$drive\"", 60);
 	    die "attaching fleecing image $volid failed - $ret\n" if $ret !~ m/OK/s;
@@ -633,6 +633,8 @@ my sub check_and_prepare_fleecing {
     }
 
     if ($use_fleecing) {
+	$self->query_block_node_sizes($vmid, $disks);
+
 	my ($default_format, $valid_formats) = PVE::Storage::storage_default_format(
 	    $self->{storecfg}, $fleecing_opts->{storage});
 	my $format = scalar(grep { $_ eq 'qcow2' } $valid_formats->@*) ? 'qcow2' : 'raw';
@@ -1038,6 +1040,31 @@ sub qga_fs_thaw {
     $self->logerr($@) if $@;
 }
 
+# The size for fleecing images needs to be exactly the same size as QEMU sees. E.g. EFI disk can bex
+# attached with a smaller size then the underyling image on the storage.
+sub query_block_node_sizes {
+    my ($self, $vmid, $disks) = @_;
+
+    my $block_info = mon_cmd($vmid, "query-block");
+    $block_info = { map { $_->{device} => $_ } $block_info->@* };
+
+    for my $diskinfo ($disks->@*) {
+	my $drive_key = $diskinfo->{virtdev};
+	$drive_key .= "-backup" if $drive_key eq 'tpmstate0';
+	my $block_node_size =
+	    eval { $block_info->{"drive-$drive_key"}->{inserted}->{image}->{'virtual-size'}; };
+	if (!$block_node_size) {
+	    $self->loginfo(
+		"could not determine block node size of drive '$drive_key' - using fallback");
+	    $block_node_size = $diskinfo->{size}
+		or die "could not determine size of drive '$drive_key'\n";
+	}
+	$diskinfo->{'block-node-size'} = $block_node_size;
+    }
+
+    return;
+}
+
 # we need a running QEMU/KVM process for backup, starts a paused (prelaunch)
 # one if VM isn't already running
 sub enforce_vm_running_for_backup {
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 19/37] backup: fleecing: use exact size when allocating non-raw fleecing images

[Fiona Ebner]

A non-1KiB aligned source image could cause issues when used with
qcow2 fleecing images, e.g. for an image with size 4.5 KiB:
> Size mismatch for 'drive-tpmstate0-backup-fleecing' - sector count 10 != 9

Raw images are attached to QEMU with an explicit 'size' argument, so
rounding up before allocation doesn't matter, but it does for qcow2.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 4a258895..6562aba6 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -558,7 +558,15 @@ my sub allocate_fleecing_images {
 		my $name = "vm-$vmid-fleece-$n";
 		$name .= ".$format" if $scfg->{path};
 
-		my $size = PVE::Tools::convert_size($di->{'block-node-size'}, 'b' => 'kb');
+		my $size;
+		if ($format ne 'raw') {
+		    # Since non-raw images cannot be attached with an explicit 'size' parameter to
+		    # QEMU later, pass the exact size to the storage layer. This makes qcow2
+		    # fleecing images work for non-1KiB-aligned source images.
+		    $size = $di->{'block-node-size'}/1024;
+		} else {
+		    $size = PVE::Tools::convert_size($di->{'block-node-size'}, 'b' => 'kb');
+		}
 
 		$di->{'fleece-volid'} = PVE::Storage::vdisk_alloc(
 		    $self->{storecfg}, $fleecing_storeid, $vmid, $format, $name, $size);
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 20/37] backup: allow adding fleecing images also for EFI and TPM

[Fiona Ebner]

For the external backup API, it will be necessary to add a fleecing
image even for small disks like EFI and TPM, because there is no other
place the old data could be copied to when a new guest write comes in.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 6562aba6..65f01791 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -541,7 +541,7 @@ my sub cleanup_fleecing_images {
 }
 
 my sub allocate_fleecing_images {
-    my ($self, $disks, $vmid, $fleecing_storeid, $format) = @_;
+    my ($self, $disks, $vmid, $fleecing_storeid, $format, $all_images) = @_;
 
     die "internal error - no fleecing storage specified\n" if !$fleecing_storeid;
 
@@ -552,7 +552,8 @@ my sub allocate_fleecing_images {
 	my $n = 0; # counter for fleecing image names
 
 	for my $di ($disks->@*) {
-	    next if $di->{virtdev} =~ m/^(?:tpmstate|efidisk)\d$/; # too small to be worth it
+	    # EFI/TPM are usually too small to be worth it, but it's required for external providers
+	    next if !$all_images && $di->{virtdev} =~ m/^(?:tpmstate|efidisk)\d$/;
 	    if ($di->{type} eq 'block' || $di->{type} eq 'file') {
 		my $scfg = PVE::Storage::storage_config($self->{storecfg}, $fleecing_storeid);
 		my $name = "vm-$vmid-fleece-$n";
@@ -624,7 +625,7 @@ my sub attach_fleecing_images {
 }
 
 my sub check_and_prepare_fleecing {
-    my ($self, $vmid, $fleecing_opts, $disks, $is_template, $qemu_support) = @_;
+    my ($self, $vmid, $fleecing_opts, $disks, $is_template, $qemu_support, $all_images) = @_;
 
     # Even if the VM was started specifically for fleecing, it's possible that the VM is resumed and
     # then starts doing IO. For VMs that are not resumed the fleecing images will just stay empty,
@@ -647,7 +648,8 @@ my sub check_and_prepare_fleecing {
 	    $self->{storecfg}, $fleecing_opts->{storage});
 	my $format = scalar(grep { $_ eq 'qcow2' } $valid_formats->@*) ? 'qcow2' : 'raw';
 
-	allocate_fleecing_images($self, $disks, $vmid, $fleecing_opts->{storage}, $format);
+	allocate_fleecing_images(
+	    $self, $disks, $vmid, $fleecing_opts->{storage}, $format, $all_images);
 	attach_fleecing_images($self, $disks, $vmid, $format);
     }
 
@@ -738,7 +740,7 @@ sub archive_pbs {
 	my $is_template = PVE::QemuConfig->is_template($self->{vmlist}->{$vmid});
 
 	$task->{'use-fleecing'} = check_and_prepare_fleecing(
-	    $self, $vmid, $opts->{fleecing}, $task->{disks}, $is_template, $qemu_support);
+	    $self, $vmid, $opts->{fleecing}, $task->{disks}, $is_template, $qemu_support, 0);
 
 	my $fs_frozen = $self->qga_fs_freeze($task, $vmid);
 
@@ -911,7 +913,7 @@ sub archive_vma {
 	$attach_tpmstate_drive->($self, $task, $vmid);
 
 	$task->{'use-fleecing'} = check_and_prepare_fleecing(
-	    $self, $vmid, $opts->{fleecing}, $task->{disks}, $is_template, $qemu_support);
+	    $self, $vmid, $opts->{fleecing}, $task->{disks}, $is_template, $qemu_support, 0);
 
 	my $outfh;
 	if ($opts->{stdout}) {
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 21/37] backup: implement backup for external providers

[Fiona Ebner]

The state of the VM's disk images at the time the backup is started is
preserved via a snapshot-access block node. Old data is moved to the
fleecing image when new guest writes come in. The snapshot-access
block node, as well as the associated bitmap in case of incremental
backup, will be made available to the external provider. They are
exported via NBD and for 'nbd' mechanism, the NBD socket path is
passed to the provider, while for 'file-handle' mechanism, the NBD
export is made accessible via a file handle and the bitmap information
is made available via a $next_dirty_region->() function. For
'file-handle', the 'nbdinfo' and 'nbdfuse' binaries are required.

The provider can indicate that it wants to do an incremental backup by
returning the bitmap ID that was used for a previous backup and it
will then be told if the bitmap was newly created (either first backup
or old bitmap was invalid) or if the bitmap can be reused.

The provider then reads the parts of the NBD or virtual file it needs,
either the full disk for full backup, or the dirty parts according to
the bitmap for incremental backup. The bitmap has to be respected,
reads to other parts of the image will return an error. After backing
up each part of the disk, it should be discarded in the export to
avoid unnecessary space usage in the fleecing image (requires the
storage underlying the fleecing image to support discard too).

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Drop outdated call to backup_get_task_size().

 PVE/VZDump/QemuServer.pm | 399 ++++++++++++++++++++++++++++++++++++++-
 1 file changed, 397 insertions(+), 2 deletions(-)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 65f01791..5e959de4 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -3,12 +3,15 @@ package PVE::VZDump::QemuServer;
 use strict;
 use warnings;
 
+use Fcntl qw(:mode);
 use File::Basename;
-use File::Path;
+use File::Path qw(make_path remove_tree);
+use File::stat qw();
 use IO::File;
 use IPC::Open3;
 use JSON;
 use POSIX qw(EINTR EAGAIN);
+use Time::HiRes qw(usleep);
 
 use PVE::Cluster qw(cfs_read_file);
 use PVE::INotify;
@@ -20,7 +23,7 @@ use PVE::QMPClient;
 use PVE::Storage::Plugin;
 use PVE::Storage::PBSPlugin;
 use PVE::Storage;
-use PVE::Tools;
+use PVE::Tools qw(run_command);
 use PVE::VZDump;
 use PVE::Format qw(render_duration render_bytes);
 
@@ -30,6 +33,7 @@ use PVE::QemuServer::Drive qw(checked_volume_format);
 use PVE::QemuServer::Helpers;
 use PVE::QemuServer::Machine;
 use PVE::QemuServer::Monitor qw(mon_cmd);
+use PVE::QemuServer::QMPHelpers;
 
 use base qw (PVE::VZDump::Plugin);
 
@@ -284,6 +288,8 @@ sub archive {
 
     if ($self->{vzdump}->{opts}->{pbs}) {
 	$self->archive_pbs($task, $vmid);
+    } elsif ($self->{vzdump}->{'backup-provider'}) {
+	$self->archive_external($task, $vmid);
     } else {
 	$self->archive_vma($task, $vmid, $filename, $comp);
     }
@@ -1148,11 +1154,90 @@ sub snapshot {
     # nothing to do
 }
 
+my sub cleanup_file_handles {
+    my ($self, $file_handles) = @_;
+
+    for my $file_handle ($file_handles->@*) {
+	close($file_handle) or $self->log('warn', "unable to close file handle - $!");
+    }
+}
+
+my sub cleanup_nbd_mounts {
+    my ($self, $info) = @_;
+
+    for my $mount_point (keys $info->%*) {
+	my $pid_file = delete($info->{$mount_point}->{'pid-file'});
+	unlink($pid_file) or $self->log('warn', "unable to unlink '$pid_file' - $!");
+	# Do a lazy unmount, because the target might still be busy even if the file handle was
+	# already closed.
+	eval { run_command(['fusermount', '-z', '-u', $mount_point ]); };
+	if (my $err = $@) {
+	    delete $info->{$mount_point};
+	    $self->log('warn', "unable to unmount NBD backup source '$mount_point' - $err");
+	}
+    }
+
+    # Wait for the unmount before cleaning up child PIDs to avoid 'nbdfuse' processes being
+    # interrupted by the signals issued there.
+    my $waited;
+    my $wait_limit = 50; # 5 seconds
+    for ($waited = 0; $waited < $wait_limit && scalar(keys $info->%*); $waited++) {
+	for my $mount_point (keys $info->%*) {
+	    delete($info->{$mount_point}) if !-e $info->{$mount_point}->{'virtual-file'};
+	    eval { remove_tree($mount_point); };
+	}
+	usleep(100_000);
+    }
+    # just informational, remaining child processes will be killed afterwards
+    $self->loginfo("unable to gracefully cleanup NBD fuse mounts") if scalar(keys $info->%*) != 0;
+}
+
+my sub cleanup_child_processes {
+    my ($self, $cpids) = @_;
+
+    my $waited;
+    my $wait_limit = 5;
+    for ($waited = 0; $waited < $wait_limit && scalar(keys $cpids->%*); $waited++) {
+	for my $cpid (keys $cpids->%*) {
+	    delete($cpids->{$cpid}) if waitpid($cpid, POSIX::WNOHANG) > 0;
+	}
+	if ($waited == 0) {
+	    kill 15, $_ for keys $cpids->%*;
+	}
+	sleep 1;
+    }
+    if ($waited == $wait_limit && scalar(keys $cpids->%*)) {
+	kill 9, $_ for keys $cpids->%*;
+	sleep 1;
+	for my $cpid (keys $cpids->%*) {
+	    delete($cpids->{$cpid}) if waitpid($cpid, POSIX::WNOHANG) > 0;
+	}
+	$self->log('warn', "unable to collect child process '$_'") for keys $cpids->%*;
+    }
+}
+
 sub cleanup {
     my ($self, $task, $vmid) = @_;
 
     # If VM was started only for backup, it is already stopped now.
     if (PVE::QemuServer::Helpers::vm_running_locally($vmid)) {
+	if ($task->{cleanup}->{'nbd-stop'}) {
+	    eval { PVE::QemuServer::QMPHelpers::nbd_stop($vmid); };
+	    $self->logerr($@) if $@;
+	}
+
+	if (my $info = $task->{cleanup}->{'backup-access-teardown'}) {
+	    my $params = {
+		'target-id' => $info->{'target-id'},
+		timeout => 60,
+		success => $info->{success} ? JSON::true : JSON::false,
+	    };
+
+	    $self->loginfo("tearing down backup-access");
+	    eval { mon_cmd($vmid, "backup-access-teardown", $params->%*) };
+	    $self->logerr($@) if $@;
+	}
+
 	$detach_tpmstate_drive->($task, $vmid);
 	detach_fleecing_images($task->{disks}, $vmid) if $task->{'use-fleecing'};
     }
@@ -1162,6 +1247,316 @@ sub cleanup {
     if ($self->{qmeventd_fh}) {
 	close($self->{qmeventd_fh});
     }
+
+    cleanup_file_handles($self, $task->{cleanup}->{'file-handles'})
+	if $task->{cleanup}->{'file-handles'};
+
+    cleanup_nbd_mounts($self, $task->{cleanup}->{'nbd-mounts'})
+	if $task->{cleanup}->{'nbd-mounts'};
+
+    cleanup_child_processes($self, $task->{cleanup}->{'child-pids'})
+	if $task->{cleanup}->{'child-pids'};
+
+    if (my $dir = $task->{'backup-access-root-dir'}) {
+	eval { remove_tree($dir) };
+	$self->log('warn', "unable to cleanup directory $dir - $@") if $@;
+    }
+}
+
+my sub virtual_file_backup_prepare {
+    my ($self, $vmid, $task, $device_name, $size, $nbd_path, $bitmap_name) = @_;
+
+    my $cleanup = $task->{cleanup};
+
+    my $nbd_uri = "nbd+unix:///${device_name}?socket=${nbd_path}";
+
+    my $error_fh;
+    my $next_dirty_region;
+
+    # If there is no dirty bitmap, it can be treated as if there's a full dirty one. The output of
+    # nbdinfo is a list of tuples with offset, length, type, description. The first bit of 'type' is
+    # set when the bitmap is dirty, see QEMU's docs/interop/nbd.txt
+    my $dirty_bitmap = [];
+    if ($bitmap_name) {
+	my $input = IO::File->new();
+	my $info = IO::File->new();
+	$error_fh = IO::File->new();
+	my $nbdinfo_cmd = ["nbdinfo", $nbd_uri, "--map=qemu:dirty-bitmap:${bitmap_name}"];
+	my $cpid = open3($input, $info, $error_fh, $nbdinfo_cmd->@*)
+	    or die "failed to spawn nbdinfo child - $!\n";
+	$cleanup->{'child-pids'}->{$cpid} = 1;
+
+	$next_dirty_region = sub {
+	    my ($offset, $length, $type);
+	    do {
+		my $line = <$info>;
+		return if !$line;
+		die "unexpected output from nbdinfo - $line\n"
+		    if $line !~ m/^\s*(\d+)\s*(\d+)\s*(\d+)/; # also untaints
+		($offset, $length, $type) = ($1, $2, $3);
+	    } while (($type & 0x1) == 0); # not dirty
+	    return ($offset, $length);
+	};
+    } else {
+	my $done = 0;
+	$next_dirty_region = sub {
+	    return if $done;
+	    $done = 1;
+	    return (0, $size);
+	};
+    }
+
+    my $mount_point = $task->{'backup-access-root-dir'}
+	."/${vmid}-nbd.backup-access.${device_name}.$$";
+    make_path($mount_point) or die "unable to create directory $mount_point\n";
+    $cleanup->{'nbd-mounts'}->{$mount_point} = {};
+
+    # Note that nbdfuse requires "$dir/$file". A single name would be treated as a dir and the file
+    # would be named "$dir/nbd" then
+    my $virtual_file = "${mount_point}/${device_name}";
+    $cleanup->{'nbd-mounts'}->{$mount_point}->{'virtual-file'} = $virtual_file;
+
+    my $pid_file = "${mount_point}.pid";
+    PVE::Tools::file_set_contents($pid_file, '', 0600);
+    $cleanup->{'nbd-mounts'}->{$mount_point}->{'pid-file'} = $pid_file;
+
+    my $cpid = fork() // die "fork failed: $!\n";
+    if (!$cpid) {
+	# By default, access will be restricted to the current user, because the allow_other fuse
+	# mount option is not used.
+	eval {
+	    run_command(
+		["nbdfuse", '--pidfile', $pid_file, $virtual_file, $nbd_uri],
+		logfunc => sub { $self->loginfo("nbdfuse '$virtual_file': $_[0]") },
+	    );
+	};
+	if (my $err = $@) {
+	    eval { $self->loginfo($err); };
+	    POSIX::_exit(1);
+	}
+	POSIX::_exit(0);
+    }
+    $cleanup->{'child-pids'}->{$cpid} = 1;
+
+    my ($virtual_file_ready, $waited) = (0, 0);
+    while (!$virtual_file_ready && $waited < 30) { # 3 seconds
+	my $pid = PVE::Tools::file_read_firstline($pid_file);
+	if ($pid) {
+	    $virtual_file_ready = 1;
+	} else {
+	    usleep(100_000);
+	    $waited++;
+	}
+    }
+    die "timeout setting up virtual file '$virtual_file'" if !$virtual_file_ready;
+
+    $self->loginfo("provided NBD export as a virtual file '$virtual_file'");
+
+    # NOTE O_DIRECT, because each block should be read exactly once and also because fuse will try
+    # to read ahead otherwise, which would produce warning messages if the next block is not
+    # mapped/allocated for the NBD export in case of incremental backup. Open as writable to support
+    # discard.
+    my $fh = IO::File->new($virtual_file, O_RDWR | O_DIRECT)
+	or die "unable to open backup source '$virtual_file' - $!\n";
+    push $cleanup->{'file-handles'}->@*, $fh;
+
+    return ($fh, $next_dirty_region);
+}
+
+my sub backup_access_to_volume_info {
+    my ($self, $vmid, $task, $backup_access_info, $mechanism, $nbd_path) = @_;
+
+    my $bitmap_action_to_status = {
+	'not-used' => 'none',
+	'not-used-removed' => 'none',
+	'new' => 'new',
+	'used' => 'reuse',
+	'invalid' => 'new',
+    };
+
+    my $volumes = {};
+
+    for my $info ($backup_access_info->@*) {
+	my $bitmap_status = 'none';
+	my $bitmap_name;
+	if (my $bitmap_action = $info->{'bitmap-action'}) {
+	    $bitmap_status = $bitmap_action_to_status->{$bitmap_action}
+		or die "got unexpected bitmap action '$bitmap_action'\n";
+
+	    $bitmap_name = $info->{'bitmap-name'} or die "bitmap-name is not present\n";
+	}
+
+	my ($device, $size) = $info->@{qw(device size)};
+
+	$volumes->{$device}->{'bitmap-mode'} = $bitmap_status;
+	$volumes->{$device}->{size} = $size;
+
+	if ($mechanism eq 'file-handle') {
+	    my ($fh, $next_dirty_region) = virtual_file_backup_prepare(
+		$self, $vmid, $task, $device, $size, $nbd_path, $bitmap_name);
+	    $volumes->{$device}->{'file-handle'} = $fh;
+	    $volumes->{$device}->{'next-dirty-region'} = $next_dirty_region;
+	} elsif ($mechanism eq 'nbd') {
+	    $volumes->{$device}->{'nbd-path'} = $nbd_path;
+	    $volumes->{$device}->{'bitmap-name'} = $bitmap_name;
+	} else {
+	    die "internal error - unkown mechanism '$mechanism'";
+	}
+    }
+
+    return $volumes;
+}
+
+sub archive_external {
+    my ($self, $task, $vmid) = @_;
+
+    $task->{'backup-access-root-dir'} = "/run/qemu-server/${vmid}.backup-access.$$/";
+    make_path($task->{'backup-access-root-dir'})
+	or die "unable to create directory $task->{'backup-access-root-dir'}\n";
+    chmod(0700, $task->{'backup-access-root-dir'})
+	or die "unable to chmod directory $task->{'backup-access-root-dir'}\n";
+
+    my $guest_config = PVE::Tools::file_get_contents("$task->{tmpdir}/qemu-server.conf");
+    my $firewall_file = "$task->{tmpdir}/qemu-server.fw";
+
+    my $opts = $self->{vzdump}->{opts};
+
+    my $backup_provider = $self->{vzdump}->{'backup-provider'};
+
+    $self->loginfo("starting external backup via " . $backup_provider->provider_name());
+
+    my $starttime = time();
+
+    # get list early so we die on unkown drive types before doing anything
+    my $devlist = _get_task_devlist($task);
+
+    $self->enforce_vm_running_for_backup($vmid);
+    $self->{qmeventd_fh} = PVE::QemuServer::register_qmeventd_handle($vmid);
+
+    eval {
+	$SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = $SIG{PIPE} = sub {
+	    die "interrupted by signal\n";
+	};
+
+	my $qemu_support = mon_cmd($vmid, "query-proxmox-support");
+
+	if (!$qemu_support->{'backup-access-api'}) {
+		die "backups access API required for external provider backup is not supported by"
+		    ." the running QEMU version. Please make sure you've installed the latest "
+		    ." version and the VM has been restarted.\n";
+	}
+
+	$attach_tpmstate_drive->($self, $task, $vmid);
+
+	my $is_template = PVE::QemuConfig->is_template($self->{vmlist}->{$vmid});
+
+	my $fleecing = check_and_prepare_fleecing(
+	    $self, $vmid, $opts->{fleecing}, $task->{disks}, $is_template, $qemu_support, 1);
+	die "cannot setup backup access without fleecing\n" if !$fleecing;
+
+	$task->{'use-fleecing'} = 1;
+
+	my $target_id = "snapshot-access:$opts->{storage}";
+
+	my ($mechanism, $bitmap_name) = $backup_provider->backup_get_mechanism($vmid, 'qemu');
+	die "mechanism '$mechanism' requested by backup provider is not supported for VMs\n"
+	    if $mechanism ne 'file-handle' && $mechanism ne 'nbd';
+
+	$self->loginfo("using backup mechanism '$mechanism'");
+
+	if ($mechanism eq 'file-handle') {
+	    # For mechanism 'file-handle', the nbdfuse binary is required. Also, the bitmap needs
+	    # to be passed to the provider. The bitmap cannot be dumped via QMP and doing it via
+	    # qemu-img is experimental, so use nbdinfo. Both are in libnbd-bin.
+	    die "need 'nbdfuse' binary from package libnbd-bin\n" if !-e "/usr/bin/nbdfuse";
+	}
+
+	my $params = {
+	    'target-id' => $target_id,
+	    devlist => $devlist,
+	    timeout => 60,
+	};
+
+	if ($bitmap_name) {
+	    # prepend storage ID so different providers can never cause clashes
+	    $bitmap_name = "$opts->{storage}-" . $bitmap_name;
+	    $params->{'bitmap-name'} = $bitmap_name;
+	}
+
+	my $fs_frozen = $self->qga_fs_freeze($task, $vmid);
+
+	$self->loginfo("setting up snapshot-access for backup");
+
+	$task->{cleanup}->{'backup-access-teardown'} = { 'target-id' => $target_id, success => 0 };
+
+	my $backup_access_info = eval { mon_cmd($vmid, "backup-access-setup", $params->%*) };
+	my $qmperr = $@;
+
+	if ($fs_frozen) {
+	    $self->qga_fs_thaw($vmid);
+	}
+
+	die $qmperr if $qmperr;
+
+	$self->resume_vm_after_job_start($task, $vmid);
+
+	my $bitmap_info = mon_cmd($vmid, 'query-pbs-bitmap-info');
+	for my $info (sort { $a->{drive} cmp $b->{drive} } $bitmap_info->@*) {
+	    my $text = $bitmap_action_to_human->($self, $info);
+	    my $drive = $info->{drive};
+	    $drive =~ s/^drive-//; # for consistency
+	    $self->loginfo("$drive: dirty-bitmap status: $text");
+	}
+
+	$self->loginfo("starting NBD server");
+
+	my $nbd_path = "$task->{'backup-access-root-dir'}/${vmid}-nbd.backup-access";
+	mon_cmd(
+	    $vmid, "nbd-server-start", addr => { type => 'unix', data => { path => $nbd_path } } );
+	$task->{cleanup}->{'nbd-stop'} = 1;
+
+	for my $info ($backup_access_info->@*) {
+	    $self->loginfo("adding NBD export for $info->{device}");
+
+	    my $export_params = {
+		id => $info->{device},
+		'node-name' => $info->{'node-name'},
+		writable => JSON::true, # for discard
+		type => "nbd",
+		name => $info->{device}, # NBD export name
+	    };
+
+	    if ($info->{'bitmap-name'}) {
+		$export_params->{bitmaps} = [{
+		    node => $info->{'bitmap-node-name'},
+		    name => $info->{'bitmap-name'},
+		}];
+	    }
+
+	    mon_cmd($vmid, "block-export-add", $export_params->%*);
+	}
+
+	my $volumes = backup_access_to_volume_info(
+	    $self, $vmid, $task, $backup_access_info, $mechanism, $nbd_path);
+
+	my $param = {};
+	$param->{'bandwidth-limit'} = $opts->{bwlimit} * 1024 if $opts->{bwlimit};
+	$param->{'firewall-config'} = PVE::Tools::file_get_contents($firewall_file)
+	    if -e $firewall_file;
+
+	$backup_provider->backup_vm($vmid, $guest_config, $volumes, $param);
+    };
+    my $err = $@;
+
+    if ($err) {
+	$self->logerr($err);
+	$self->resume_vm_after_job_start($task, $vmid);
+    } else {
+	$task->{cleanup}->{'backup-access-teardown'}->{success} = 1;
+    }
+    $self->restore_vm_power_state($vmid);
+
+    die $err if $err;
 }
 
 1;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 22/37] test: qemu img convert: add test cases for snapshots

[Fiona Ebner]

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 test/run_qemu_img_convert_tests.pl | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/test/run_qemu_img_convert_tests.pl b/test/run_qemu_img_convert_tests.pl
index 20ff387a..29c188d8 100755
--- a/test/run_qemu_img_convert_tests.pl
+++ b/test/run_qemu_img_convert_tests.pl
@@ -194,6 +194,24 @@ my $tests = [
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw",
 	]
     },
+    {
+	name => "lvmsnapshot",
+	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, 'foo', 0, undef ],
+	expected => [
+	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
+	    "/dev/pve/snap_vm-$vmid-disk-0_foo",
+	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw",
+	]
+    },
+    {
+	name => "qcow2snapshot",
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.qcow2", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, 'snap', 0, undef ],
+	expected => [
+	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-l", "snapshot.name=snap", "-f", "qcow2", "-O", "raw",
+	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.qcow2",
+	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw",
+	]
+    },
 ];
 
 my $command;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 23/37] image convert: collect options in hash argument

[Fiona Ebner]

In preparation to add another option and to improve style for the
callers.

One of the test cases that specified $is_zero_initialized is for a
non-existent storage, so the option was not added there.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 PVE/QemuServer.pm                  | 19 +++++++---
 PVE/QemuServer/ImportDisk.pm       |  3 +-
 test/run_qemu_img_convert_tests.pl | 58 ++++++++++++++++++++----------
 3 files changed, 56 insertions(+), 24 deletions(-)

diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index ffd5d56b..5c6cb94a 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -7787,8 +7787,14 @@ sub convert_iscsi_path {
     die "cannot convert iscsi path '$path', unknown format\n";
 }
 
+# The possible options are:
+# bwlimit - The bandwidth limit in KiB/s.
+# is-zero-initialized - If the destination image is zero-initialized.
+# snapname - Use this snapshot of the source image.
 sub qemu_img_convert {
-    my ($src_volid, $dst_volid, $size, $snapname, $is_zero_initialized, $bwlimit) = @_;
+    my ($src_volid, $dst_volid, $size, $opts) = @_;
+
+    my ($bwlimit, $snapname) = $opts->@{qw(bwlimit snapname)};
 
     my $storecfg = PVE::Storage::config();
     my ($src_storeid) = PVE::Storage::parse_volume_id($src_volid, 1);
@@ -7846,7 +7852,7 @@ sub qemu_img_convert {
 
     push @$cmd, $src_path;
 
-    if (!$dst_is_iscsi && $is_zero_initialized) {
+    if (!$dst_is_iscsi && $opts->{'is-zero-initialized'}) {
 	push @$cmd, "zeroinit:$dst_path";
     } else {
 	push @$cmd, $dst_path;
@@ -8307,7 +8313,12 @@ sub clone_disk {
 		push $cmd->@*, "bs=$bs", "osize=$size", "if=$src_path", "of=$dst_path";
 		run_command($cmd);
 	    } else {
-		qemu_img_convert($drive->{file}, $newvolid, $size, $snapname, $sparseinit, $bwlimit);
+		my $opts = {
+		    bwlimit => $bwlimit,
+		    'is-zero-initialized' => $sparseinit,
+		    snapname => $snapname,
+		};
+		qemu_img_convert($drive->{file}, $newvolid, $size, $opts);
 	    }
 	}
     }
@@ -8391,7 +8402,7 @@ sub create_efidisk($$$$$$$) {
     my $volid = PVE::Storage::vdisk_alloc($storecfg, $storeid, $vmid, $fmt, undef, $vars_size);
     PVE::Storage::activate_volumes($storecfg, [$volid]);
 
-    qemu_img_convert($ovmf_vars, $volid, $vars_size_b, undef, 0);
+    qemu_img_convert($ovmf_vars, $volid, $vars_size_b);
     my $size = PVE::Storage::volume_size_info($storecfg, $volid, 3);
 
     return ($volid, $size/1024);
diff --git a/PVE/QemuServer/ImportDisk.pm b/PVE/QemuServer/ImportDisk.pm
index 30d56bae..75b5e0de 100755
--- a/PVE/QemuServer/ImportDisk.pm
+++ b/PVE/QemuServer/ImportDisk.pm
@@ -70,7 +70,8 @@ sub do_import {
 	    local $SIG{PIPE} = sub { die "interrupted by signal $!\n"; };
 
 	PVE::Storage::activate_volumes($storecfg, [$dst_volid]);
-	PVE::QemuServer::qemu_img_convert($src_path, $dst_volid, $src_size, undef, $zeroinit);
+	PVE::QemuServer::qemu_img_convert(
+	    $src_path, $dst_volid, $src_size, { 'is-zero-initialized' => $zeroinit });
 	PVE::Storage::deactivate_volumes($storecfg, [$dst_volid]);
 	PVE::QemuConfig->lock_config($vmid, $create_drive) if !$params->{'skip-config-update'};
     };
diff --git a/test/run_qemu_img_convert_tests.pl b/test/run_qemu_img_convert_tests.pl
index 29c188d8..652e61f2 100755
--- a/test/run_qemu_img_convert_tests.pl
+++ b/test/run_qemu_img_convert_tests.pl
@@ -55,7 +55,7 @@ my $storage_config = {
 my $tests = [
     {
 	name => 'qcow2raw',
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.qcow2", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.qcow2", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "qcow2", "-O", "raw",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.qcow2", "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw"
@@ -63,7 +63,7 @@ my $tests = [
     },
     {
 	name => "raw2qcow2",
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "local:$vmid/vm-$vmid-disk-0.qcow2", 1024*10, undef, 0, undef ],
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "local:$vmid/vm-$vmid-disk-0.qcow2", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "qcow2",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw", "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.qcow2"
@@ -71,7 +71,7 @@ my $tests = [
     },
     {
 	name => "local2rbd",
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "rbd-store:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "rbd-store:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw", "rbd:cpool/vm-$vmid-disk-0:mon_host=127.0.0.42;127.0.0.21;[\\:\\:1]:auth_supported=none"
@@ -79,7 +79,7 @@ my $tests = [
     },
     {
 	name => "rbd2local",
-	parameters => [ "rbd-store:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "rbd-store:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "rbd:cpool/vm-$vmid-disk-0:mon_host=127.0.0.42;127.0.0.21;[\\:\\:1]:auth_supported=none", "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw"
@@ -87,7 +87,7 @@ my $tests = [
     },
     {
 	name => "local2zos",
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "--target-image-opts",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw",
@@ -96,7 +96,7 @@ my $tests = [
     },
     {
 	name => "zos2local",
-	parameters => [ "zfs-over-iscsi:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "zfs-over-iscsi:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "--image-opts", "-O", "raw",
 	    "file.driver=iscsi,file.transport=tcp,file.initiator-name=foobar,file.portal=127.0.0.1,file.target=iqn.2019-10.org.test:foobar,file.lun=1,driver=raw",
@@ -105,7 +105,7 @@ my $tests = [
     },
     {
 	name => "zos2rbd",
-	parameters => [ "zfs-over-iscsi:vm-$vmid-disk-0", "rbd-store:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "zfs-over-iscsi:vm-$vmid-disk-0", "rbd-store:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "--image-opts", "-O", "raw",
 	    "file.driver=iscsi,file.transport=tcp,file.initiator-name=foobar,file.portal=127.0.0.1,file.target=iqn.2019-10.org.test:foobar,file.lun=1,driver=raw",
@@ -114,7 +114,7 @@ my $tests = [
     },
     {
 	name => "rbd2zos",
-	parameters => [ "rbd-store:vm-$vmid-disk-0", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "rbd-store:vm-$vmid-disk-0", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "--target-image-opts",
 	    "rbd:cpool/vm-$vmid-disk-0:mon_host=127.0.0.42;127.0.0.21;[\\:\\:1]:auth_supported=none",
@@ -123,7 +123,7 @@ my $tests = [
     },
     {
 	name => "local2lvmthin",
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "local-lvm:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "local:$vmid/vm-$vmid-disk-0.raw", "local-lvm:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.raw",
@@ -132,7 +132,7 @@ my $tests = [
     },
     {
 	name => "lvmthin2local",
-	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "/dev/pve/vm-$vmid-disk-0",
@@ -141,7 +141,12 @@ my $tests = [
     },
     {
 	name => "zeroinit",
-	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 1, undef ],
+	parameters => [
+	    "local-lvm:vm-$vmid-disk-0",
+	    "local:$vmid/vm-$vmid-disk-0.raw",
+	    1024*10,
+	    { 'is-zero-initialized' => 1 },
+	],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "/dev/pve/vm-$vmid-disk-0",
@@ -150,12 +155,12 @@ my $tests = [
     },
     {
 	name => "notexistingstorage",
-	parameters => [ "local-lvm:vm-$vmid-disk-0", "not-existing:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 1, undef ],
+	parameters => [ "local-lvm:vm-$vmid-disk-0", "not-existing:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => "storage 'not-existing' does not exist\n",
     },
     {
 	name => "vmdkfile",
-	parameters => [ "./test.vmdk", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "./test.vmdk", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "vmdk", "-O", "raw",
 	    "./test.vmdk",
@@ -164,12 +169,12 @@ my $tests = [
     },
     {
 	name => "notexistingfile",
-	parameters => [ "/foo/bar", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "/foo/bar", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => "source '/foo/bar' is not a valid volid nor path for qemu-img convert\n",
     },
     {
 	name => "efidisk",
-	parameters => [ "/usr/share/kvm/OVMF_VARS-pure-efi.fd", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, undef ],
+	parameters => [ "/usr/share/kvm/OVMF_VARS-pure-efi.fd", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-O", "raw",
 	    "/usr/share/kvm/OVMF_VARS-pure-efi.fd",
@@ -178,7 +183,7 @@ my $tests = [
     },
     {
 	name => "efi2zos",
-	parameters => [ "/usr/share/kvm/OVMF_VARS-pure-efi.fd", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10, undef, 0, undef ],
+	parameters => [ "/usr/share/kvm/OVMF_VARS-pure-efi.fd", "zfs-over-iscsi:vm-$vmid-disk-0", 1024*10 ],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "--target-image-opts",
 	    "/usr/share/kvm/OVMF_VARS-pure-efi.fd",
@@ -187,7 +192,12 @@ my $tests = [
     },
     {
 	name => "bwlimit",
-	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, undef, 0, 1024 ],
+	parameters => [
+	    "local-lvm:vm-$vmid-disk-0",
+	    "local:$vmid/vm-$vmid-disk-0.raw",
+	    1024*10,
+	    { bwlimit => 1024 },
+	],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-r", "1024K", "-f", "raw", "-O", "raw",
 	    "/dev/pve/vm-$vmid-disk-0",
@@ -196,7 +206,12 @@ my $tests = [
     },
     {
 	name => "lvmsnapshot",
-	parameters => [ "local-lvm:vm-$vmid-disk-0", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, 'foo', 0, undef ],
+	parameters => [
+	    "local-lvm:vm-$vmid-disk-0",
+	    "local:$vmid/vm-$vmid-disk-0.raw",
+	    1024*10,
+	    { snapname => 'foo'},
+	],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-f", "raw", "-O", "raw",
 	    "/dev/pve/snap_vm-$vmid-disk-0_foo",
@@ -205,7 +220,12 @@ my $tests = [
     },
     {
 	name => "qcow2snapshot",
-	parameters => [ "local:$vmid/vm-$vmid-disk-0.qcow2", "local:$vmid/vm-$vmid-disk-0.raw", 1024*10, 'snap', 0, undef ],
+	parameters => [
+	    "local:$vmid/vm-$vmid-disk-0.qcow2",
+	    "local:$vmid/vm-$vmid-disk-0.raw",
+	    1024*10,
+	    { snapname => 'snap' },
+	],
 	expected => [
 	    "/usr/bin/qemu-img", "convert", "-p", "-n", "-l", "snapshot.name=snap", "-f", "qcow2", "-O", "raw",
 	    "/var/lib/vz/images/$vmid/vm-$vmid-disk-0.qcow2",
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 24/37] image convert: allow caller to specify the format of the source path

[Fiona Ebner]

In preparation for the restore API for backup providers that doesn't
want detection based on the file extension but always requires raw.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

New in v6.

 PVE/QemuServer.pm | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 5c6cb94a..93f985be 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -7791,6 +7791,8 @@ sub convert_iscsi_path {
 # bwlimit - The bandwidth limit in KiB/s.
 # is-zero-initialized - If the destination image is zero-initialized.
 # snapname - Use this snapshot of the source image.
+# source-path-format - Indicate the format of the source when the source is a path. For PVE-managed
+# volumes, the format from the storage layer is always used.
 sub qemu_img_convert {
     my ($src_volid, $dst_volid, $size, $opts) = @_;
 
@@ -7816,7 +7818,9 @@ sub qemu_img_convert {
 	$cachemode = 'none' if $src_scfg->{type} eq 'zfspool';
     } elsif (-f $src_volid || -b $src_volid) {
 	$src_path = $src_volid;
-	if ($src_path =~ m/\.($PVE::QemuServer::Drive::QEMU_FORMAT_RE)$/) {
+	if ($opts->{'source-path-format'}) {
+	    $src_format = $opts->{'source-path-format'};
+	} elsif ($src_path =~ m/\.($PVE::QemuServer::Drive::QEMU_FORMAT_RE)$/) {
 	    $src_format = $1;
 	}
     }
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 25/37] backup: implement restore for external providers

[Fiona Ebner]

First, the provider is asked about what restore mechanism to use.
Currently, only 'qemu-img' is possible. Then the configuration files
are restored, the provider gives information about volumes contained
in the backup and finally the volumes are restored via
'qemu-img convert'.

The code for the restore_external_archive() function was copied and
adapted from the restore_proxmox_backup_archive() function. Together
with restore_vma_archive() it seems sensible to extract the common
parts and use a dedicated module for restore code.

The parse_restore_archive() helper was renamed, because it's not just
parsing.

While currently, the format for the source can only be raw, do an
untrusted check for the source for future-proofing. Still serves as a
basic sanity check currently.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Drop superfluous $storeid argument for $backup_provider->restore_*
  method calls.
* Pass raw format to 'qemu-img convert' for source.
* Use zeroinit filter when target volume supports the 'sparseinit'
  feature.
* Squash together with patch with untrusted check.
* Adapt to renamed archive_get_*_config() methods.

 PVE/API2/Qemu.pm  |  30 +++++++++-
 PVE/QemuServer.pm | 149 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 176 insertions(+), 3 deletions(-)

diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm
index 156b1c7b..6c7c1d0d 100644
--- a/PVE/API2/Qemu.pm
+++ b/PVE/API2/Qemu.pm
@@ -927,7 +927,7 @@ __PACKAGE__->register_method({
 	return $res;
     }});
 
-my $parse_restore_archive = sub {
+my $classify_restore_archive = sub {
     my ($storecfg, $archive) = @_;
 
     my ($archive_storeid, $archive_volname) = PVE::Storage::parse_volume_id($archive, 1);
@@ -941,6 +941,22 @@ my $parse_restore_archive = sub {
 	    $res->{type} = 'pbs';
 	    return $res;
 	}
+	if (PVE::Storage::storage_has_feature($storecfg, $archive_storeid, 'backup-provider')) {
+	    my $log_function = sub {
+		my ($log_level, $message) = @_;
+		my $prefix = $log_level eq 'err' ? 'ERROR' : uc($log_level);
+		print "$prefix: $message\n";
+	    };
+	    my $backup_provider = PVE::Storage::new_backup_provider(
+		$storecfg,
+		$archive_storeid,
+		$log_function,
+	    );
+
+	    $res->{type} = 'external';
+	    $res->{'backup-provider'} = $backup_provider;
+	    return $res;
+	}
     }
     my $path = PVE::Storage::abs_filesystem_path($storecfg, $archive);
     $res->{type} = 'file';
@@ -1101,7 +1117,7 @@ __PACKAGE__->register_method({
 		    'backup',
 		);
 
-		$archive = $parse_restore_archive->($storecfg, $archive);
+		$archive = $classify_restore_archive->($storecfg, $archive);
 	    }
 	}
 
@@ -1160,7 +1176,15 @@ __PACKAGE__->register_method({
 			PVE::QemuServer::check_restore_permissions($rpcenv, $authuser, $merged);
 		    }
 		}
-		if ($archive->{type} eq 'file' || $archive->{type} eq 'pipe') {
+		if (my $backup_provider = $archive->{'backup-provider'}) {
+		    PVE::QemuServer::restore_external_archive(
+			$backup_provider,
+			$archive->{volid},
+			$vmid,
+			$authuser,
+			$restore_options,
+		    );
+		} elsif ($archive->{type} eq 'file' || $archive->{type} eq 'pipe') {
 		    die "live-restore is only compatible with backup images from a Proxmox Backup Server\n"
 			if $live_restore;
 		    PVE::QemuServer::restore_file_archive($archive->{path} // '-', $vmid, $authuser, $restore_options);
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 93f985be..b4796010 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -7175,6 +7175,155 @@ sub restore_proxmox_backup_archive {
     }
 }
 
+sub restore_external_archive {
+    my ($backup_provider, $archive, $vmid, $user, $options) = @_;
+
+    die "live restore from backup provider is not implemented\n" if $options->{live};
+
+    my $storecfg = PVE::Storage::config();
+
+    my ($storeid, $volname) = PVE::Storage::parse_volume_id($archive);
+    my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
+
+    my $tmpdir = "/run/qemu-server/vzdumptmp$$";
+    rmtree($tmpdir);
+    mkpath($tmpdir) or die "unable to create $tmpdir\n";
+
+    my $conffile = PVE::QemuConfig->config_file($vmid);
+    # disable interrupts (always do cleanups)
+    local $SIG{INT} =
+	local $SIG{TERM} =
+	local $SIG{QUIT} =
+	local $SIG{HUP} = sub { print STDERR "got interrupt - ignored\n"; };
+
+    # Note: $oldconf is undef if VM does not exists
+    my $cfs_path = PVE::QemuConfig->cfs_config_path($vmid);
+    my $oldconf = PVE::Cluster::cfs_read_file($cfs_path);
+    my $new_conf_raw = '';
+
+    my $rpcenv = PVE::RPCEnvironment::get();
+    my $devinfo = {}; # info about drives included in backup
+    my $virtdev_hash = {}; # info about allocated drives
+
+    eval {
+	# enable interrupts
+	local $SIG{INT} =
+	    local $SIG{TERM} =
+	    local $SIG{QUIT} =
+	    local $SIG{HUP} =
+	    local $SIG{PIPE} = sub { die "interrupted by signal\n"; };
+
+	my $cfgfn = "$tmpdir/qemu-server.conf";
+	my $firewall_config_fn = "$tmpdir/fw.conf";
+
+	my $cmd = "restore";
+
+	my ($mechanism, $vmtype) =
+	    $backup_provider->restore_get_mechanism($volname);
+	die "mechanism '$mechanism' requested by backup provider is not supported for VMs\n"
+	    if $mechanism ne 'qemu-img';
+	die "cannot restore non-VM guest of type '$vmtype'\n" if $vmtype ne 'qemu';
+
+	$devinfo = $backup_provider->restore_vm_init($volname);
+
+	my $data = $backup_provider->archive_get_guest_config($volname)
+	    or die "backup provider failed to extract guest configuration\n";
+	PVE::Tools::file_set_contents($cfgfn, $data);
+
+	if ($data = $backup_provider->archive_get_firewall_config($volname)) {
+	    PVE::Tools::file_set_contents($firewall_config_fn, $data);
+	    my $pve_firewall_dir = '/etc/pve/firewall';
+	    mkdir $pve_firewall_dir; # make sure the dir exists
+	    PVE::Tools::file_copy($firewall_config_fn, "${pve_firewall_dir}/$vmid.fw");
+	}
+
+	my $fh = IO::File->new($cfgfn, "r") or die "unable to read qemu-server.conf - $!\n";
+
+	$virtdev_hash = $parse_backup_hints->($rpcenv, $user, $storecfg, $fh, $devinfo, $options);
+
+	# create empty/temp config
+	PVE::Tools::file_set_contents($conffile, "memory: 128\nlock: create");
+
+	$restore_cleanup_oldconf->($storecfg, $vmid, $oldconf, $virtdev_hash) if $oldconf;
+
+	# allocate volumes
+	my $map = $restore_allocate_devices->($storecfg, $virtdev_hash, $vmid);
+
+	for my $virtdev (sort keys $virtdev_hash->%*) {
+	    my $d = $virtdev_hash->{$virtdev};
+	    next if $d->{is_cloudinit}; # no need to restore cloudinit
+
+	    my $sparseinit = PVE::Storage::volume_has_feature($storecfg, 'sparseinit', $d->{volid});
+	    my $source_format = 'raw';
+
+	    my $info = $backup_provider->restore_vm_volume_init($volname, $d->{devname}, {});
+	    my $source_path = $info->{'qemu-img-path'}
+		or die "did not get source image path from backup provider\n";
+
+	    print "importing drive '$d->{devname}' from '$source_path'\n";
+
+	    # safety check for untrusted source image
+	    PVE::Storage::file_size_info($source_path, undef, $source_format, 1);
+
+	    eval {
+		my $convert_opts = {
+		    bwlimit => $options->{bwlimiit},
+		    'is-zero-initialized' => $sparseinit,
+		    'source-path-format' => $source_format,
+		};
+		qemu_img_convert($source_path, $d->{volid}, $d->{size}, $convert_opts);
+	    };
+	    my $err = $@;
+	    eval { $backup_provider->restore_vm_volume_cleanup($volname, $d->{devname}, {}); };
+	    if (my $cleanup_err = $@) {
+		die $cleanup_err if !$err;
+		warn $cleanup_err;
+	    }
+	    die $err if $err
+	}
+
+	$fh->seek(0, 0) || die "seek failed - $!\n";
+
+	my $cookie = { netcount => 0 };
+	while (defined(my $line = <$fh>)) {
+	    $new_conf_raw .= restore_update_config_line(
+		$cookie,
+		$map,
+		$line,
+		$options->{unique},
+	    );
+	}
+
+	$fh->close();
+    };
+    my $err = $@;
+
+    eval { $backup_provider->restore_vm_cleanup($volname); };
+    warn "backup provider cleanup after restore failed - $@" if $@;
+
+    if ($err) {
+	$restore_deactivate_volumes->($storecfg, $virtdev_hash);
+    }
+
+    rmtree($tmpdir);
+
+    if ($err) {
+	$restore_destroy_volumes->($storecfg, $virtdev_hash);
+	die $err;
+    }
+
+    my $new_conf = restore_merge_config($conffile, $new_conf_raw, $options->{override_conf});
+    check_restore_permissions($rpcenv, $user, $new_conf);
+    PVE::QemuConfig->write_config($vmid, $new_conf);
+
+    eval { rescan($vmid, 1); };
+    warn $@ if $@;
+
+    PVE::AccessControl::add_vm_to_pool($vmid, $options->{pool}) if $options->{pool};
+
+    return;
+}
+
 sub pbs_live_restore {
     my ($vmid, $conf, $storecfg, $restored_disks, $opts) = @_;
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 26/37] backup: future-proof checks for QEMU feature support

[Fiona Ebner]

The features returned by the 'query-proxmox-support' QMP command are
booleans, so just checking for definedness is not enough in principle.
In practice, a feature is currently always true if defined. Still, fix
the checks, should the need to disable support for a feature ever
arise in the future and to avoid propagating the pattern further.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 5e959de4..0d4949a8 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -639,7 +639,7 @@ my sub check_and_prepare_fleecing {
 
     my $use_fleecing = $fleecing_opts && $fleecing_opts->{enabled} && !$is_template;
 
-    if ($use_fleecing && !defined($qemu_support->{'backup-fleecing'})) {
+    if ($use_fleecing && !$qemu_support->{'backup-fleecing'}) {
 	$self->log(
 	    'warn',
 	    "running QEMU version does not support backup fleecing - continuing without",
@@ -739,7 +739,7 @@ sub archive_pbs {
 
 	# pve-qemu supports it since 5.2.0-1 (PVE 6.4), so safe to die since PVE 8
 	die "master key configured but running QEMU version does not support master keys\n"
-	    if !defined($qemu_support->{'pbs-masterkey'}) && defined($master_keyfile);
+	    if !$qemu_support->{'pbs-masterkey'} && defined($master_keyfile);
 
 	$attach_tpmstate_drive->($self, $task, $vmid);
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 27/37] backup: support 'missing-recreated' bitmap action

[Fiona Ebner]

A new 'missing-recreated' action was added on the QEMU side.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 0d4949a8..02e2c5af 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -316,6 +316,8 @@ my $bitmap_action_to_human = sub {
 	}
     } elsif ($action eq "invalid") {
 	return "existing bitmap was invalid and has been cleared";
+    } elsif ($action eq "missing-recreated") {
+	return "expected bitmap was missing and has been recreated";
     } else {
 	return "unknown";
     }
@@ -1372,6 +1374,7 @@ my sub backup_access_to_volume_info {
 	'new' => 'new',
 	'used' => 'reuse',
 	'invalid' => 'new',
+	'missing-recreated' => 'new',
     };
 
     my $volumes = {};
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH qemu-server v6 28/37] backup: bitmap action to human: lie about TPM state

[Fiona Ebner]

The TPM state drive is newly attached each time, so it is fully
expected that a bitmap from last time would be missing.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 PVE/VZDump/QemuServer.pm | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/PVE/VZDump/QemuServer.pm b/PVE/VZDump/QemuServer.pm
index 02e2c5af..726f128f 100644
--- a/PVE/VZDump/QemuServer.pm
+++ b/PVE/VZDump/QemuServer.pm
@@ -317,6 +317,8 @@ my $bitmap_action_to_human = sub {
     } elsif ($action eq "invalid") {
 	return "existing bitmap was invalid and has been cleared";
     } elsif ($action eq "missing-recreated") {
+	# Lie about the TPM state, because it is newly attached each time.
+	return "created new" if $info->{drive} eq 'drive-tpmstate0-backup';
 	return "expected bitmap was missing and has been recreated";
     } else {
 	return "unknown";
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 29/37] add LXC::Namespaces module

[Fiona Ebner]

The module includes a run_in_userns() helper to run a Perl subroutine
in a user namespace.

The first use case is running the container backup subroutine for
external providers inside a user namespace. That allows them to see
the filesystem to back-up from the containers perspective and also
improves security because of isolation.

Heavily adapted from code by Wolfgang from the pve-buildpkg
repository.

Originally-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
[FE: add $idmap parameter, drop $aux_groups parameter
     use different fork helper
     use newuidmap and newgidmap binaries]
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* import constants from PVE::Tools instead of redefining.

 src/PVE/LXC/Makefile      |  1 +
 src/PVE/LXC/Namespaces.pm | 57 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 58 insertions(+)
 create mode 100644 src/PVE/LXC/Namespaces.pm

diff --git a/src/PVE/LXC/Makefile b/src/PVE/LXC/Makefile
index a190260..4b05e98 100644
--- a/src/PVE/LXC/Makefile
+++ b/src/PVE/LXC/Makefile
@@ -5,6 +5,7 @@ SOURCES= \
 	Create.pm \
 	Migrate.pm \
 	Monitor.pm \
+	Namespaces.pm \
 	Setup.pm \
 	Tools.pm
 
diff --git a/src/PVE/LXC/Namespaces.pm b/src/PVE/LXC/Namespaces.pm
new file mode 100644
index 0000000..bd39c07
--- /dev/null
+++ b/src/PVE/LXC/Namespaces.pm
@@ -0,0 +1,57 @@
+package PVE::LXC::Namespaces;
+
+use strict;
+use warnings;
+
+use Fcntl qw(O_WRONLY);
+use Socket;
+
+use PVE::Tools qw(CLONE_NEWNS CLONE_NEWUSER);
+
+my sub set_id_map($$) {
+    my ($pid, $id_map) = @_;
+
+    my @gid_args = ();
+    my @uid_args = ();
+
+    for my $map ($id_map->@*) {
+	my ($type, $ct, $host, $length) = $map->@*;
+
+	push @gid_args, $ct, $host, $length if $type eq 'g';
+	push @uid_args, $ct, $host, $length if $type eq 'u';
+    }
+
+    PVE::Tools::run_command(['newgidmap', $pid, @gid_args]) if scalar(@gid_args);
+    PVE::Tools::run_command(['newuidmap', $pid, @uid_args]) if scalar(@uid_args);
+}
+
+sub run_in_userns(&;$) {
+    my ($code, $id_map) = @_;
+    socketpair(my $sp, my $sc, AF_UNIX, SOCK_STREAM, PF_UNSPEC)
+	or die "socketpair: $!\n";
+    my $child = sub {
+	close($sp);
+	PVE::Tools::unshare(CLONE_NEWUSER|CLONE_NEWNS) or die "unshare(NEWUSER|NEWNS): $!\n";
+	syswrite($sc, "1\n") == 2 or die "write: $!\n";
+	shutdown($sc, 1);
+	my $two = <$sc>;
+	die "failed to sync with parent process\n" if $two ne "2\n";
+	close($sc);
+	$! = undef;
+	($(, $)) = (0, 0); die "$!\n" if $!;
+	($<, $>) = (0, 0); die "$!\n" if $!;
+	return $code->();
+    };
+    my $parent = sub {
+	my ($pid) = @_;
+	close($sc);
+	my $one = <$sp>;
+	die "failed to sync with userprocess\n" if $one ne "1\n";
+	set_id_map($pid, $id_map);
+	syswrite($sp, "2\n") == 2 or die "write: $!\n";
+	close($sp);
+    };
+    PVE::Tools::run_fork($child, { afterfork => $parent });
+}
+
+1;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 30/37] backup: implement backup for external providers

[Fiona Ebner]

The filesystem structure is made available as a directory in a
consistent manner (with details depending on the vzdump backup mode)
just like for regular backup via tar.

The backup_container() method of the backup provider is executed in
a user namespace with the container's ID mapping applied. This allows
the backup provider to see the container's filesystem from the
container's perspective.

The 'prepare' phase of the backup hook is executed right before and
allows the backup provider to prepare for the (usually) unprivileged
execution context in the user namespace.

The backup provider needs to back up the guest and firewall
configuration and the filesystem structure of the container, honoring
file exclusions and the bandwidth limit.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Switch to new backup_container_prepare() method.

 src/PVE/VZDump/LXC.pm | 40 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 39 insertions(+), 1 deletion(-)

diff --git a/src/PVE/VZDump/LXC.pm b/src/PVE/VZDump/LXC.pm
index 9029387..79d3a31 100644
--- a/src/PVE/VZDump/LXC.pm
+++ b/src/PVE/VZDump/LXC.pm
@@ -11,6 +11,7 @@ use PVE::Cluster qw(cfs_read_file);
 use PVE::GuestHelpers;
 use PVE::INotify;
 use PVE::LXC::Config;
+use PVE::LXC::Namespaces;
 use PVE::LXC;
 use PVE::Storage;
 use PVE::Tools;
@@ -124,6 +125,7 @@ sub prepare {
 
     my ($id_map, $root_uid, $root_gid) = PVE::LXC::parse_id_maps($conf);
     $task->{userns_cmd} = PVE::LXC::userns_command($id_map);
+    $task->{id_map} = $id_map;
     $task->{root_uid} = $root_uid;
     $task->{root_gid} = $root_gid;
 
@@ -373,7 +375,43 @@ sub archive {
     my $userns_cmd = $task->{userns_cmd};
     my $findexcl = $self->{vzdump}->{findexcl};
 
-    if ($self->{vzdump}->{opts}->{pbs}) {
+    if (my $backup_provider = $self->{vzdump}->{'backup-provider'}) {
+	$self->loginfo("starting external backup via " . $backup_provider->provider_name());
+
+	if (!scalar($task->{id_map}->@*) || $task->{root_uid} == 0 || $task->{root_gid} == 0) {
+	    $self->log("warn", "external backup of privileged container can only be restored as"
+		." unprivileged which might not work in all cases");
+	}
+
+	my ($mechanism) = $backup_provider->backup_get_mechanism($vmid, 'lxc');
+	die "mechanism '$mechanism' requested by backup provider is not supported for containers\n"
+	    if $mechanism ne 'directory';
+
+	$self->loginfo("using backup mechanism '$mechanism'");
+
+	my $guest_config = PVE::Tools::file_get_contents("$tmpdir/etc/vzdump/pct.conf");
+	my $firewall_file = "$tmpdir/etc/vzdump/pct.fw";
+
+	my $info = {
+	    directory => $snapdir,
+	    sources => [@sources],
+	    'backup-user-id' => $task->{root_uid},
+	};
+	$info->{'firewall-config'} = PVE::Tools::file_get_contents($firewall_file)
+	    if -e $firewall_file;
+	$info->{'bandwidth-limit'} = $opts->{bwlimit} * 1024 if $opts->{bwlimit};
+
+	$backup_provider->backup_container_prepare($vmid, $info);
+
+	if (scalar($task->{id_map}->@*)) {
+	    PVE::LXC::Namespaces::run_in_userns(
+		sub { $backup_provider->backup_container($vmid, $guest_config, $findexcl, $info); },
+		$task->{id_map},
+	    );
+	} else {
+	    $backup_provider->backup_container($vmid, $guest_config, $findexcl, $info);
+	}
+    } elsif ($self->{vzdump}->{opts}->{pbs}) {
 
 	my $param = [];
 	push @$param, "pct.conf:$tmpdir/etc/vzdump/pct.conf";
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 31/37] backup: implement restore for external providers

[Fiona Ebner]

First, the provider is asked about what restore mechanism to use.
Currently, 'directory' and 'tar' are possible. The 'directory'
mechanism is for restoring from a directory containing the container's
full filesystem structure, which is restored by piping from a
privileged tar cf - to tar xf - in the associated user namespace. The
'tar' mechanism is for restoring a (potentially compressed) tar file
containing the container's full filesystem structure.

The new functions are copied and adapted from the existing ones for
PBS or tar and it might be worth to factor out the common parts.

Restore of containers as privileged are prohibited, because the
archives from an external provider are considered less trusted than
from Proxmox VE storages. If ever allowing that in the future, at
least it would be worth extracting the tar archive in a restricted
context (e.g. user namespace with ID mapped mount or seccomp).

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Adapt to renamed archive_get_*_config() methods.

 src/PVE/LXC/Create.pm | 149 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 149 insertions(+)

diff --git a/src/PVE/LXC/Create.pm b/src/PVE/LXC/Create.pm
index 8c8cb9a..c3c7640 100644
--- a/src/PVE/LXC/Create.pm
+++ b/src/PVE/LXC/Create.pm
@@ -7,6 +7,7 @@ use File::Path;
 use Fcntl;
 
 use PVE::RPCEnvironment;
+use PVE::RESTEnvironment qw(log_warn);
 use PVE::Storage::PBSPlugin;
 use PVE::Storage::Plugin;
 use PVE::Storage;
@@ -26,6 +27,24 @@ sub restore_archive {
 	if ($scfg->{type} eq 'pbs') {
 	    return restore_proxmox_backup_archive($storage_cfg, $archive, $rootdir, $conf, $no_unpack_error, $bwlimit);
 	}
+	if (PVE::Storage::storage_has_feature($storage_cfg, $storeid, 'backup-provider')) {
+	    my $log_function = sub {
+		my ($log_level, $message) = @_;
+		my $prefix = $log_level eq 'err' ? 'ERROR' : uc($log_level);
+		print "$prefix: $message\n";
+	    };
+	    my $backup_provider =
+		PVE::Storage::new_backup_provider($storage_cfg, $storeid, $log_function);
+	    return restore_external_archive(
+		$backup_provider,
+		$storeid,
+		$volname,
+		$rootdir,
+		$conf,
+		$no_unpack_error,
+		$bwlimit,
+	    );
+	}
     }
 
     $archive = PVE::Storage::abs_filesystem_path($storage_cfg, $archive) if $archive ne '-';
@@ -127,6 +146,62 @@ sub restore_tar_archive {
     die $err if $err && !$no_unpack_error;
 }
 
+sub restore_external_archive {
+    my ($backup_provider, $storeid, $volname, $rootdir, $conf, $no_unpack_error, $bwlimit) = @_;
+
+    die "refusing to restore privileged container backup from external source\n"
+	if !$conf->{unprivileged};
+
+    my ($mechanism, $vmtype) = $backup_provider->restore_get_mechanism($volname, $storeid);
+    die "cannot restore non-LXC guest of type '$vmtype'\n" if $vmtype ne 'lxc';
+
+    my $info = $backup_provider->restore_container_init($volname, $storeid, {});
+    eval {
+	if ($mechanism eq 'tar') {
+	    my $tar_path = $info->{'tar-path'}
+		or die "did not get path to tar file from backup provider\n";
+	    die "not a regular file '$tar_path'" if !-f $tar_path;
+	    restore_tar_archive($tar_path, $rootdir, $conf, $no_unpack_error, $bwlimit);
+	} elsif ($mechanism eq 'directory') {
+	    my $directory = $info->{'archive-directory'}
+		or die "did not get path to archive directory from backup provider\n";
+	    die "not a directory '$directory'" if !-d $directory;
+
+	    my $create_cmd = [
+		'tar',
+		'cpf',
+		'-',
+		@PVE::Storage::Plugin::COMMON_TAR_FLAGS,
+		"--directory=$directory",
+		'.',
+	    ];
+
+	    my $extract_cmd = restore_tar_archive_command($conf, undef, $rootdir, $bwlimit);
+
+	    my $cmd;
+	    # if there is a bandwidth limit, the command is already a nested array reference
+	    if (ref($extract_cmd) eq 'ARRAY' && ref($extract_cmd->[0]) eq 'ARRAY') {
+		$cmd = [$create_cmd, $extract_cmd->@*];
+	    } else {
+		$cmd = [$create_cmd, $extract_cmd];
+	    }
+
+	    eval { PVE::Tools::run_command($cmd); };
+	    die $@ if $@ && !$no_unpack_error;
+	} else {
+	    die "mechanism '$mechanism' requested by backup provider is not supported for LXCs\n";
+	}
+    };
+    my $err = $@;
+    eval { $backup_provider->restore_container_cleanup($volname, $storeid, {}); };
+    if (my $cleanup_err = $@) {
+	die $cleanup_err if !$err;
+	warn $cleanup_err;
+    }
+    die $err if $err;
+
+}
+
 sub recover_config {
     my ($storage_cfg, $volid, $vmid) = @_;
 
@@ -135,6 +210,8 @@ sub recover_config {
 	my $scfg = PVE::Storage::storage_check_enabled($storage_cfg, $storeid);
 	if ($scfg->{type} eq 'pbs') {
 	    return recover_config_from_proxmox_backup($storage_cfg, $volid, $vmid);
+	} elsif (PVE::Storage::storage_has_feature($storage_cfg, $storeid, 'backup-provider')) {
+	    return recover_config_from_external_backup($storage_cfg, $volid, $vmid);
 	}
     }
 
@@ -209,6 +286,26 @@ sub recover_config_from_tar {
     return wantarray ? ($conf, $mp_param) : $conf;
 }
 
+sub recover_config_from_external_backup {
+    my ($storage_cfg, $volid, $vmid) = @_;
+
+    $vmid //= 0;
+
+    my $raw = PVE::Storage::extract_vzdump_config($storage_cfg, $volid);
+
+    my $conf = PVE::LXC::Config::parse_pct_config("/lxc/${vmid}.conf" , $raw);
+
+    delete $conf->{snapshots};
+
+    my $mp_param = {};
+    PVE::LXC::Config->foreach_volume($conf, sub {
+	my ($ms, $mountpoint) = @_;
+	$mp_param->{$ms} = $conf->{$ms};
+    });
+
+    return wantarray ? ($conf, $mp_param) : $conf;
+}
+
 sub restore_configuration {
     my ($vmid, $storage_cfg, $archive, $rootdir, $conf, $restricted, $unique, $skip_fw) = @_;
 
@@ -218,6 +315,26 @@ sub restore_configuration {
 	if ($scfg->{type} eq 'pbs') {
 	    return restore_configuration_from_proxmox_backup($vmid, $storage_cfg, $archive, $rootdir, $conf, $restricted, $unique, $skip_fw);
 	}
+	if (PVE::Storage::storage_has_feature($storage_cfg, $storeid, 'backup-provider')) {
+	    my $log_function = sub {
+		my ($log_level, $message) = @_;
+		my $prefix = $log_level eq 'err' ? 'ERROR' : uc($log_level);
+		print "$prefix: $message\n";
+	    };
+	    my $backup_provider =
+		PVE::Storage::new_backup_provider($storage_cfg, $storeid, $log_function);
+	    return restore_configuration_from_external_backup(
+		$backup_provider,
+		$vmid,
+		$storage_cfg,
+		$archive,
+		$rootdir,
+		$conf,
+		$restricted,
+		$unique,
+		$skip_fw,
+	    );
+	}
     }
     restore_configuration_from_etc_vzdump($vmid, $rootdir, $conf, $restricted, $unique, $skip_fw);
 }
@@ -258,6 +375,38 @@ sub restore_configuration_from_proxmox_backup {
     }
 }
 
+sub restore_configuration_from_external_backup {
+    my ($backup_provider, $vmid, $storage_cfg, $archive, $rootdir, $conf, $restricted, $unique, $skip_fw) = @_;
+
+    my ($storeid, $volname) = PVE::Storage::parse_volume_id($archive);
+    my $scfg = PVE::Storage::storage_config($storage_cfg, $storeid);
+
+    my ($vtype, $name, undef, undef, undef, undef, $format) =
+	PVE::Storage::parse_volname($storage_cfg, $archive);
+
+    my $oldconf = recover_config_from_external_backup($storage_cfg, $archive, $vmid);
+
+    sanitize_and_merge_config($conf, $oldconf, $restricted, $unique);
+
+    my $firewall_config =
+	$backup_provider->archive_get_firewall_config($volname, $storeid);
+
+    if ($firewall_config) {
+	my $pve_firewall_dir = '/etc/pve/firewall';
+	my $pct_fwcfg_target = "${pve_firewall_dir}/${vmid}.fw";
+	if ($skip_fw) {
+	    warn "ignoring firewall config from backup archive, lacking API permission to modify firewall.\n";
+	    warn "old firewall configuration in '$pct_fwcfg_target' left in place!\n"
+		if -e $pct_fwcfg_target;
+	} else {
+	    mkdir $pve_firewall_dir; # make sure the directory exists
+	    PVE::Tools::file_set_contents($pct_fwcfg_target, $firewall_config);
+	}
+    }
+
+    return;
+}
+
 sub sanitize_and_merge_config {
     my ($conf, $oldconf, $restricted, $unique) = @_;
 
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 32/37] external restore: don't use 'one-file-system' tar flag when restoring from a directory

[Fiona Ebner]

This gives backup providers more freedom, e.g. mount backed-up mount
point volumes individually.

Suggested-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/LXC/Create.pm | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/PVE/LXC/Create.pm b/src/PVE/LXC/Create.pm
index c3c7640..d0cbb1e 100644
--- a/src/PVE/LXC/Create.pm
+++ b/src/PVE/LXC/Create.pm
@@ -167,11 +167,15 @@ sub restore_external_archive {
 		or die "did not get path to archive directory from backup provider\n";
 	    die "not a directory '$directory'" if !-d $directory;
 
+	    # Give backup provider more freedom, e.g. mount backed-up mount point volumes
+	    # individually.
+	    my @flags = grep { $_ ne '--one-file-system' } @PVE::Storage::Plugin::COMMON_TAR_FLAGS;
+
 	    my $create_cmd = [
 		'tar',
 		'cpf',
 		'-',
-		@PVE::Storage::Plugin::COMMON_TAR_FLAGS,
+		@flags,
 		"--directory=$directory",
 		'.',
 	    ];
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 33/37] create: factor out compression option helper

[Fiona Ebner]

In preparation to re-use it for checking potentially untrusted
archives.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/LXC/Create.pm | 51 +++++++++++++++++++++++++------------------
 1 file changed, 30 insertions(+), 21 deletions(-)

diff --git a/src/PVE/LXC/Create.pm b/src/PVE/LXC/Create.pm
index d0cbb1e..43fc5fe 100644
--- a/src/PVE/LXC/Create.pm
+++ b/src/PVE/LXC/Create.pm
@@ -78,15 +78,38 @@ sub restore_proxmox_backup_archive {
 	$scfg, $storeid, $cmd, $param, userns_cmd => $userns_cmd);
 }
 
+my sub tar_compression_option {
+    my ($archive) = @_;
+
+    my %compression_map = (
+	'.gz'  => '-z',
+	'.bz2' => '-j',
+	'.xz'  => '-J',
+	'.lzo'  => '--lzop',
+	'.zst'  => '--zstd',
+    );
+    if ($archive =~ /\.tar(\.[^.]+)?$/) {
+	if (defined($1)) {
+	    die "unrecognized compression format: $1\n" if !defined($compression_map{$1});
+	    return $compression_map{$1};
+	}
+	return;
+    } else {
+	die "file does not look like a template archive: $archive\n";
+    }
+}
+
 my sub restore_tar_archive_command {
-    my ($conf, $opts, $rootdir, $bwlimit) = @_;
+    my ($conf, $compression_opt, $rootdir, $bwlimit) = @_;
 
     my ($id_map, $root_uid, $root_gid) = PVE::LXC::parse_id_maps($conf);
     my $userns_cmd = PVE::LXC::userns_command($id_map);
 
-    my $cmd = [@$userns_cmd, 'tar', 'xpf', '-', $opts->@*, '--totals',
-               @PVE::Storage::Plugin::COMMON_TAR_FLAGS,
-               '-C', $rootdir];
+    my $cmd = [@$userns_cmd, 'tar', 'xpf', '-'];
+    push $cmd->@*, $compression_opt if $compression_opt;
+    push $cmd->@*, '--totals';
+    push $cmd->@*, @PVE::Storage::Plugin::COMMON_TAR_FLAGS;
+    push $cmd->@*, '-C', $rootdir;
 
     # skip-old-files doesn't have anything to do with time (old/new), but is
     # simply -k (annoyingly also called --keep-old-files) without the 'treat
@@ -108,24 +131,10 @@ sub restore_tar_archive {
 
     my $archive_fh;
     my $tar_input = '<&STDIN';
-    my @compression_opt;
+    my $compression_opt;
     if ($archive ne '-') {
 	# GNU tar refuses to autodetect this... *sigh*
-	my %compression_map = (
-	    '.gz'  => '-z',
-	    '.bz2' => '-j',
-	    '.xz'  => '-J',
-	    '.lzo'  => '--lzop',
-	    '.zst'  => '--zstd',
-	);
-	if ($archive =~ /\.tar(\.[^.]+)?$/) {
-	    if (defined($1)) {
-		die "unrecognized compression format: $1\n" if !defined($compression_map{$1});
-		@compression_opt = $compression_map{$1};
-	    }
-	} else {
-	    die "file does not look like a template archive: $archive\n";
-	}
+	$compression_opt = tar_compression_option($archive);
 	sysopen($archive_fh, $archive, O_RDONLY)
 	    or die "failed to open '$archive': $!\n";
 	my $flags = $archive_fh->fcntl(Fcntl::F_GETFD(), 0);
@@ -133,7 +142,7 @@ sub restore_tar_archive {
 	$tar_input = '<&'.fileno($archive_fh);
     }
 
-    my $cmd = restore_tar_archive_command($conf, [@compression_opt], $rootdir, $bwlimit);
+    my $cmd = restore_tar_archive_command($conf, $compression_opt, $rootdir, $bwlimit);
 
     if ($archive eq '-') {
 	print "extracting archive from STDIN\n";
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 34/37] restore tar archive: check potentially untrusted archive

[Fiona Ebner]

'tar' itself already protects against '..' in component names and
strips absolute member names when extracting (if not used with the
--absolute-names option) and in general seems sane for extracting.
Additionally, the extraction already happens in the user namespace
associated to the container. So for now, start out with some basic
sanity checks. The helper can still be extended with more checks.

Checks:

* list files in archive - will already catch many corrupted/bogus
  archives.

* check that there are at least 10 members - should also catch
  archives not actually containing a container root filesystem or
  structural issues early.

* check that /sbin directory or link exists in archive - ideally the
  check would be for /sbin/init, but this cannot be done efficiently
  before extraction, because it would require to keep track of the
  whole archive to be able to follow symlinks.

* abort if there is a multi-volume member in the archive - cheap and
  is never expected.

Checks that were considered, but not (yet) added:

* abort when a file has unrealistically large size - while this could
  help to detect certain kinds of bogus archives, there can be valid.
  use cases for extremely large sparse files, so it's not clear what
  a good limit would be (1 EiB maybe?). Also, an attacker could just
  adapt to such a limit creating multiple files and the actual
  extraction is already limited by the size of the allocated container
  volume.

* check that /sbin/init exists after extracting - cannot be done
  efficiently before extraction, because it would require to keep
  track of the whole archive to be able to follow symlinks. During
  setup there already is detection of /etc/os-release, so issues with
  the structure will already be noticed. Adding a hard fail for
  untrusted archives would require either passing that information to
  the setup phase or extracting the protected_call method from there
  into a helper.

* adding 'restrict' to the (common) tar flags - the tar manual (not
  the man page) documents: "Disable use of some potentially harmful
  'tar' options.  Currently this option disables shell invocation from
  multi-volume menu.". The flag was introduced in 2005 and this is
  still the only thing it is used for. Trying to restore a
  multi-volume archive already fails without giving multiple '--file'
  arguments and '--multi-volume', so don't bother adding the flag.

* check format of tar file - would require yet another invocation of
  the decompressor and there seems to be no built-in way to just
  display the format with 'tar'. The 'file' program could be used, but
  it seems to not make a distinction between old GNU and GNU and old
  POSIX and POSIX formats, with the old ones being candidates to
  prohibit. So that would leave just detecting the old 'v7' format.

Suggested-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/LXC/Create.pm | 75 ++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 71 insertions(+), 4 deletions(-)

diff --git a/src/PVE/LXC/Create.pm b/src/PVE/LXC/Create.pm
index 43fc5fe..53c584b 100644
--- a/src/PVE/LXC/Create.pm
+++ b/src/PVE/LXC/Create.pm
@@ -99,12 +99,73 @@ my sub tar_compression_option {
     }
 }
 
+# Basic checks trying to detect issues with a potentially untrusted or bogus tar archive.
+# Just listing the files is already a good check against corruption.
+# 'tar' itself already protects against '..' in component names and strips absolute member names
+# when extracting, so no need to check for those here.
+my sub check_tar_archive {
+    my ($archive) = @_;
+
+    print "checking archive..\n";
+
+    # To resolve links to get to 'sbin/init' would mean keeping track of everything in the archive,
+    # because the target might be ordered first. Check only that 'sbin' exists here.
+    my $found_sbin;
+
+    # Just to detect bogus archives, any valid container filesystem should have more than this.
+    my $required_members = 10;
+    my $member_count = 0;
+
+    my $check_file_list = sub {
+	my ($line) = @_;
+
+	$member_count++;
+
+	# Not always just a single number, e.g. for character devices.
+	my $size_re = qr/\d+(?:,\d+)?/;
+
+	# The date is in ISO 8601 format. The last part contains the potentially quoted file name,
+	# potentially followed by some additional info (e.g. where a link points to).
+	my ($type, $perms, $uid, $gid, $size, $date, $time, $file_info) =
+	    $line =~ m!^([a-zA-Z\-])(\S+)\s+(\d+)/(\d+)\s+($size_re)\s+(\S+)\s+(\S+)\s+(.*)$!;
+	if (!defined($type)) {
+	    print "check tar: unable to parse line: $line\n";
+	    return;
+	}
+
+	die "found multi-volume member in archive\n" if $type eq 'M';
+
+	if (!$found_sbin && (
+	    ($file_info =~ m!^(?:\./)?sbin/$! && $type eq 'd')
+	    || ($file_info =~ m!^(?:\./)?sbin ->! && $type eq 'l')
+	    || ($file_info =~ m!^(?:\./)?sbin link to! && $type eq 'h')
+	)) {
+	    $found_sbin = 1;
+	}
+
+    };
+
+    my $compression_opt = tar_compression_option($archive);
+
+    my $cmd = ['tar', '-tvf', $archive];
+    push $cmd->@*, $compression_opt if $compression_opt;
+    push $cmd->@*, '--numeric-owner';
+
+    PVE::Tools::run_command($cmd, outfunc => $check_file_list);
+
+    die "no 'sbin' directory (or link) found in archive '$archive'\n" if !$found_sbin;
+    die "less than 10 members in archive '$archive'\n" if $member_count < $required_members;
+}
+
 my sub restore_tar_archive_command {
-    my ($conf, $compression_opt, $rootdir, $bwlimit) = @_;
+    my ($conf, $compression_opt, $rootdir, $bwlimit, $untrusted) = @_;
 
     my ($id_map, $root_uid, $root_gid) = PVE::LXC::parse_id_maps($conf);
     my $userns_cmd = PVE::LXC::userns_command($id_map);
 
+    die "refusing to restore privileged container backup from external source\n"
+	if $untrusted && ($root_uid == 0 || $root_gid == 0);
+
     my $cmd = [@$userns_cmd, 'tar', 'xpf', '-'];
     push $cmd->@*, $compression_opt if $compression_opt;
     push $cmd->@*, '--totals';
@@ -127,7 +188,7 @@ my sub restore_tar_archive_command {
 }
 
 sub restore_tar_archive {
-    my ($archive, $rootdir, $conf, $no_unpack_error, $bwlimit) = @_;
+    my ($archive, $rootdir, $conf, $no_unpack_error, $bwlimit, $untrusted) = @_;
 
     my $archive_fh;
     my $tar_input = '<&STDIN';
@@ -142,7 +203,12 @@ sub restore_tar_archive {
 	$tar_input = '<&'.fileno($archive_fh);
     }
 
-    my $cmd = restore_tar_archive_command($conf, $compression_opt, $rootdir, $bwlimit);
+    if ($untrusted) {
+	die "cannot verify untrusted archive on STDIN\n" if $archive eq '-';
+	check_tar_archive($archive);
+    }
+
+    my $cmd = restore_tar_archive_command($conf, $compression_opt, $rootdir, $bwlimit, $untrusted);
 
     if ($archive eq '-') {
 	print "extracting archive from STDIN\n";
@@ -170,7 +236,7 @@ sub restore_external_archive {
 	    my $tar_path = $info->{'tar-path'}
 		or die "did not get path to tar file from backup provider\n";
 	    die "not a regular file '$tar_path'" if !-f $tar_path;
-	    restore_tar_archive($tar_path, $rootdir, $conf, $no_unpack_error, $bwlimit);
+	    restore_tar_archive($tar_path, $rootdir, $conf, $no_unpack_error, $bwlimit, 1);
 	} elsif ($mechanism eq 'directory') {
 	    my $directory = $info->{'archive-directory'}
 		or die "did not get path to archive directory from backup provider\n";
@@ -189,6 +255,7 @@ sub restore_external_archive {
 		'.',
 	    ];
 
+	    # archive is trusted, we created it
 	    my $extract_cmd = restore_tar_archive_command($conf, undef, $rootdir, $bwlimit);
 
 	    my $cmd;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH container v6 35/37] api: add early check against restoring privileged container from external source

[Fiona Ebner]

While restore_external_archive() already has a check, that happens
after an existing container is destroyed.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/API2/LXC.pm | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm
index 7cb5122..6cd771c 100644
--- a/src/PVE/API2/LXC.pm
+++ b/src/PVE/API2/LXC.pm
@@ -40,6 +40,17 @@ BEGIN {
     }
 }
 
+my sub assert_not_restore_from_external {
+    my ($archive, $storage_cfg) = @_;
+
+    my ($storeid, undef) = PVE::Storage::parse_volume_id($archive, 1);
+
+    return if !defined($storeid);
+    return if !PVE::Storage::storage_has_feature($storage_cfg, $storeid, 'backup-provider');
+
+    die "refusing to restore privileged container backup from external source\n";
+}
+
 my $check_storage_access_migrate = sub {
     my ($rpcenv, $authuser, $storecfg, $storage, $node) = @_;
 
@@ -409,6 +420,9 @@ __PACKAGE__->register_method({
 			$conf->{unprivileged} = $orig_conf->{unprivileged}
 			    if !defined($unprivileged) && defined($orig_conf->{unprivileged});
 
+			assert_not_restore_from_external($archive, $storage_cfg)
+			    if !$conf->{unprivileged};
+
 			# implicit privileged change is checked here
 			if ($old_conf->{unprivileged} && !$conf->{unprivileged}) {
 			    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Allocate']);
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH manager v6 36/37] ui: backup: also check for backup subtype to classify archive

[Fiona Ebner]

In anticipation of future storage plugins that might not have
PBS-specific formats or adhere to the vzdump naming scheme for
backups.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 www/manager6/Utils.js              | 10 ++++++----
 www/manager6/grid/BackupView.js    |  4 ++--
 www/manager6/storage/BackupView.js |  4 ++--
 3 files changed, 10 insertions(+), 8 deletions(-)

diff --git a/www/manager6/Utils.js b/www/manager6/Utils.js
index aa415759..1ed3de5d 100644
--- a/www/manager6/Utils.js
+++ b/www/manager6/Utils.js
@@ -694,12 +694,14 @@ Ext.define('PVE.Utils', {
 	'import': gettext('Import'),
     },
 
-    volume_is_qemu_backup: function(volid, format) {
-	return format === 'pbs-vm' || volid.match(':backup/vzdump-qemu-');
+    volume_is_qemu_backup: function(volume) {
+	return volume.format === 'pbs-vm' || volume.volid.match(':backup/vzdump-qemu-') ||
+	    volume.subtype === 'qemu';
     },
 
-    volume_is_lxc_backup: function(volid, format) {
-	return format === 'pbs-ct' || volid.match(':backup/vzdump-(lxc|openvz)-');
+    volume_is_lxc_backup: function(volume) {
+	return volume.format === 'pbs-ct' || volume.volid.match(':backup/vzdump-(lxc|openvz)-') ||
+	    volume.subtype === 'lxc';
     },
 
     authSchema: {
diff --git a/www/manager6/grid/BackupView.js b/www/manager6/grid/BackupView.js
index e71d1c88..ef3649c6 100644
--- a/www/manager6/grid/BackupView.js
+++ b/www/manager6/grid/BackupView.js
@@ -29,11 +29,11 @@ Ext.define('PVE.grid.BackupView', {
 	var vmtypeFilter;
 	if (vmtype === 'lxc' || vmtype === 'openvz') {
 	    vmtypeFilter = function(item) {
-		return PVE.Utils.volume_is_lxc_backup(item.data.volid, item.data.format);
+		return PVE.Utils.volume_is_lxc_backup(item.data);
 	    };
 	} else if (vmtype === 'qemu') {
 	    vmtypeFilter = function(item) {
-		return PVE.Utils.volume_is_qemu_backup(item.data.volid, item.data.format);
+		return PVE.Utils.volume_is_qemu_backup(item.data);
 	    };
 	} else {
 	    throw "unsupported VM type '" + vmtype + "'";
diff --git a/www/manager6/storage/BackupView.js b/www/manager6/storage/BackupView.js
index db184def..749c2136 100644
--- a/www/manager6/storage/BackupView.js
+++ b/www/manager6/storage/BackupView.js
@@ -86,9 +86,9 @@ Ext.define('PVE.storage.BackupView', {
 		disabled: true,
 		handler: function(b, e, rec) {
 		    let vmtype;
-		    if (PVE.Utils.volume_is_qemu_backup(rec.data.volid, rec.data.format)) {
+		    if (PVE.Utils.volume_is_qemu_backup(rec.data)) {
 			vmtype = 'qemu';
-		    } else if (PVE.Utils.volume_is_lxc_backup(rec.data.volid, rec.data.format)) {
+		    } else if (PVE.Utils.volume_is_lxc_backup(rec.data)) {
 			vmtype = 'lxc';
 		    } else {
 			return;
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [PATCH manager v6 37/37] backup: implement backup for external providers

[Fiona Ebner]

Call job_{init,cleanup}() and backup_{init,cleanup}() methods so that
backup providers can prepare and clean up for the whole backup job and
for individual guest backups.

It is necessary to adapt some log messages and special case some
things like is already done for PBS, e.g. log file handling.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---

Changes in v6:
* Replace job/backup hooks with dedicated methods.
* Remove backup_get_archive_name() method. The archive name is now
  returned as part of the backup_init() method. That was also a bit of
  a misnomer, since it was necessary for the backup provider to
  remember that archive name for later.
* Supersede backup_get_task_size() method and return stats as part of
  the backup_cleanup() method (in the success case). Currently, this
  only includes the archive size as well, but could be extended later
  to allow custom stats to include in the task notification summary.

 PVE/VZDump.pm           | 60 ++++++++++++++++++++++++++++++++++++-----
 test/vzdump_new_test.pl |  3 +++
 2 files changed, 56 insertions(+), 7 deletions(-)

diff --git a/PVE/VZDump.pm b/PVE/VZDump.pm
index fd89945e..82914e56 100644
--- a/PVE/VZDump.pm
+++ b/PVE/VZDump.pm
@@ -217,7 +217,10 @@ sub storage_info {
     $info->{'prune-backups'} = PVE::JSONSchema::parse_property_string('prune-backups', $scfg->{'prune-backups'})
 	if defined($scfg->{'prune-backups'});
 
-    if ($type eq 'pbs') {
+    if (PVE::Storage::storage_has_feature($cfg, $storage, 'backup-provider')) {
+	$info->{'backup-provider'} =
+	    PVE::Storage::new_backup_provider($cfg, $storage, sub { debugmsg($_[0], $_[1]); });
+    } elsif ($type eq 'pbs') {
 	$info->{pbs} = 1;
     } else {
 	$info->{dumpdir} = PVE::Storage::get_backup_dir($cfg, $storage);
@@ -717,6 +720,7 @@ sub new {
 	    $opts->{scfg} = $info->{scfg};
 	    $opts->{pbs} = $info->{pbs};
 	    $opts->{'prune-backups'} //= $info->{'prune-backups'};
+	    $self->{'backup-provider'} = $info->{'backup-provider'} if $info->{'backup-provider'};
 	}
     } elsif ($opts->{dumpdir}) {
 	$add_error->("dumpdir '$opts->{dumpdir}' does not exist")
@@ -1001,7 +1005,7 @@ sub exec_backup_task {
 	    }
 	}
 
-	if (!$self->{opts}->{pbs}) {
+	if (!$self->{opts}->{pbs} && !$self->{'backup-provider'}) {
 	    $task->{logfile} = "$opts->{dumpdir}/$basename.log";
 	}
 
@@ -1011,7 +1015,10 @@ sub exec_backup_task {
 	    $ext .= ".${comp_ext}";
 	}
 
-	if ($self->{opts}->{pbs}) {
+	if ($self->{'backup-provider'}) {
+	    die "unable to pipe backup to stdout\n" if $opts->{stdout};
+	    # the archive name $task->{target} is returned by the start hook a bit later
+	} elsif ($self->{opts}->{pbs}) {
 	    die "unable to pipe backup to stdout\n" if $opts->{stdout};
 	    $task->{target} = $pbs_snapshot_name;
 	} else {
@@ -1029,7 +1036,7 @@ sub exec_backup_task {
 	my $pid = $$;
 	if ($opts->{tmpdir}) {
 	    $task->{tmpdir} = "$opts->{tmpdir}/vzdumptmp${pid}_$vmid/";
-	} elsif ($self->{opts}->{pbs}) {
+	} elsif ($self->{opts}->{pbs} || $self->{'backup-provider'}) {
 	    $task->{tmpdir} = "/var/tmp/vzdumptmp${pid}_$vmid";
 	} else {
 	    # dumpdir is posix? then use it as temporary dir
@@ -1101,6 +1108,11 @@ sub exec_backup_task {
 	if ($mode eq 'stop') {
 	    $plugin->prepare ($task, $vmid, $mode);
 
+	    if ($self->{'backup-provider'}) {
+		my $init_result =
+		    $self->{'backup-provider'}->backup_init($vmid, $vmtype, $task->{backup_time});
+		$task->{target} = $init_result->{'archive-name'};
+	    }
 	    $self->run_hook_script ('backup-start', $task, $logfd);
 
 	    if ($running) {
@@ -1115,6 +1127,11 @@ sub exec_backup_task {
 	} elsif ($mode eq 'suspend') {
 	    $plugin->prepare ($task, $vmid, $mode);
 
+	    if ($self->{'backup-provider'}) {
+		my $init_result =
+		    $self->{'backup-provider'}->backup_init($vmid, $vmtype, $task->{backup_time});
+		$task->{target} = $init_result->{'archive-name'};
+	    }
 	    $self->run_hook_script ('backup-start', $task, $logfd);
 
 	    if ($vmtype eq 'lxc') {
@@ -1141,6 +1158,11 @@ sub exec_backup_task {
 	    }
 
 	} elsif ($mode eq 'snapshot') {
+	    if ($self->{'backup-provider'}) {
+		my $init_result =
+		    $self->{'backup-provider'}->backup_init($vmid, $vmtype, $task->{backup_time});
+		$task->{target} = $init_result->{'archive-name'};
+	    }
 	    $self->run_hook_script ('backup-start', $task, $logfd);
 
 	    my $snapshot_count = $task->{snapshot_count} || 0;
@@ -1183,11 +1205,13 @@ sub exec_backup_task {
 	    return;
 	}
 
-	my $archive_txt = $self->{opts}->{pbs} ? 'Proxmox Backup Server' : 'vzdump';
+	my $archive_txt = 'vzdump';
+	$archive_txt = 'Proxmox Backup Server' if $self->{opts}->{pbs};
+	$archive_txt = $self->{'backup-provider'}->provider_name() if $self->{'backup-provider'};
 	debugmsg('info', "creating $archive_txt archive '$task->{target}'", $logfd);
 	$plugin->archive($task, $vmid, $task->{tmptar}, $comp);
 
-	if ($self->{opts}->{pbs}) {
+	if ($self->{'backup-provider'} || $self->{opts}->{pbs}) {
 	    # size is added to task struct in guest vzdump plugins
 	} else {
 	    rename ($task->{tmptar}, $task->{target}) ||
@@ -1201,7 +1225,8 @@ sub exec_backup_task {
 
 	# Mark as protected before pruning.
 	if (my $storeid = $opts->{storage}) {
-	    my $volname = $opts->{pbs} ? $task->{target} : basename($task->{target});
+	    my $volname = $opts->{pbs} || $self->{'backup-provider'} ? $task->{target}
+	                                                             : basename($task->{target});
 	    my $volid = "${storeid}:backup/${volname}";
 
 	    if ($opts->{'notes-template'} && $opts->{'notes-template'} ne '') {
@@ -1254,6 +1279,10 @@ sub exec_backup_task {
 	    debugmsg ('info', "pruned $pruned backup(s)${log_pruned_extra}", $logfd);
 	}
 
+	if ($self->{'backup-provider'}) {
+	    my $cleanup_result = $self->{'backup-provider'}->backup_cleanup($vmid, $vmtype, 1, {});
+	    $task->{size} = $cleanup_result->{stats}->{'archive-size'};
+	}
 	$self->run_hook_script ('backup-end', $task, $logfd);
     };
     my $err = $@;
@@ -1313,6 +1342,14 @@ sub exec_backup_task {
 	debugmsg ('err', "Backup of VM $vmid failed - $err", $logfd, 1);
 	debugmsg ('info', "Failed at " . strftime("%F %H:%M:%S", localtime()));
 
+	if ($self->{'backup-provider'}) {
+	    eval {
+		$self->{'backup-provider'}->backup_cleanup(
+		    $vmid, $task->{vmtype}, 0, { error => $err });
+	    };
+	    debugmsg('warn', "backup cleanup for external provider failed - $@") if $@;
+	}
+
 	eval { $self->run_hook_script ('backup-abort', $task, $logfd); };
 	debugmsg('warn', $@) if $@; # message already contains command with phase name
 
@@ -1340,6 +1377,8 @@ sub exec_backup_task {
 		};
 		debugmsg('warn', "$@") if $@; # $@ contains already error prefix
 	    }
+	} elsif ($self->{'backup-provider'}) {
+	    $self->{'backup-provider'}->backup_handle_log_file($vmid, $task->{tmplog});
 	} elsif ($task->{logfile}) {
 	    system {'cp'} 'cp', $task->{tmplog}, $task->{logfile};
 	}
@@ -1398,6 +1437,7 @@ sub exec_backup {
     my $errcount = 0;
     eval {
 
+	$self->{'backup-provider'}->job_init($starttime) if $self->{'backup-provider'};
 	$self->run_hook_script ('job-start', undef, $job_start_fd);
 
 	foreach my $task (@$tasklist) {
@@ -1405,11 +1445,17 @@ sub exec_backup {
 	    $errcount += 1 if $task->{state} ne 'ok';
 	}
 
+	$self->{'backup-provider'}->job_cleanup() if $self->{'backup-provider'};
 	$self->run_hook_script ('job-end', undef, $job_end_fd);
     };
     my $err = $@;
 
     if ($err) {
+	if ($self->{'backup-provider'}) {
+	    eval { $self->{'backup-provider'}->job_cleanup(); };
+	    $err .= "job cleanup for external provider failed - $@" if $@;
+	}
+
 	eval { $self->run_hook_script ('job-abort', undef, $job_end_fd); };
 	$err .= $@ if $@;
 	debugmsg ('err', "Backup job failed - $err", undef, 1);
diff --git a/test/vzdump_new_test.pl b/test/vzdump_new_test.pl
index 8cd73075..01f2a661 100755
--- a/test/vzdump_new_test.pl
+++ b/test/vzdump_new_test.pl
@@ -51,6 +51,9 @@ $pve_storage_module->mock(
     activate_storage => sub {
 	return;
     },
+    get_backup_provider => sub {
+	return;
+    },
 );
 
 my $pve_cluster_module = Test::MockModule->new('PVE::Cluster');
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] applied: [PATCH common v6 09/37] syscall: expose fallocate syscall

[Thomas Lamprecht]

Am 31.03.25 um 15:19 schrieb Fiona Ebner:
> Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
> ---
>  src/PVE/Syscall.pm | 1 +
>  1 file changed, 1 insertion(+)
> 
>

applied this one already, thanks!


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Re: [pve-devel] [PATCH-SERIES qemu/common/storage/qemu-server/container/manager v6 00/37] backup provider API

[Fiona Ebner]

Am 31.03.25 um 15:19 schrieb Fiona Ebner:
> The backup provider API is split into two parts, both of which again
> need different implementations for VM and LXC guests:
> 
> 1. Backup API

Forgot to change the cover letter here. The updated description of the
API for v6 can be found here:
https://lore.proxmox.com/pve-devel/20250331132020.105324-13-f.ebner@proxmox.com/


_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

[pve-devel] [FOLLOWUP storage] backup provider: base: document limitation of backup_container() method

[Fiona Ebner]

The backup_container() method is called after forking, so changes to
$self will not be visible in later method calls done by the parent.

Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
 src/PVE/BackupProvider/Plugin/Base.pm | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/src/PVE/BackupProvider/Plugin/Base.pm b/src/PVE/BackupProvider/Plugin/Base.pm
index 424feb7..4a35330 100644
--- a/src/PVE/BackupProvider/Plugin/Base.pm
+++ b/src/PVE/BackupProvider/Plugin/Base.pm
@@ -595,11 +595,16 @@ that is made available for the mechanism returned by C<backup_get_mechanism>.
 Returns when done backing up. Ideally, the method should log the progress during
 backup.
 
-Note that this function is executed as the ID-mapped root user of the container,
+Note that this method is executed as the ID-mapped root user of the container,
 so a potentially unprivileged user. The ID is passed along as part of C<$info>.
 Use the C<backup_container_prepare()> method for preparation. For example, to
 make credentials available to the potentially unprivileged user.
 
+Note that changes to C<$self> made during this method will not be visible in
+later method calls. This is because the method is executed in a separate
+execution context after forking. Use the C<backup_container_prepare()> method
+if you need persistent changes to C<$self>.
+
 Parameters:
 
 =over
-- 
2.39.5



_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel