From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate001.proxmox.com (gate001.proxmox.com [IPv6:2a0f:8001:1:32::40]) by lore.proxmox.com (Postfix) with ESMTPS id D50381FF0DA for ; Thu, 19 Nov 2026 20:20:42 +0100 (CET) Received: from gate001.proxmox.com (localhost.localdomain [127.0.0.1]) by gate001.proxmox.com (Proxmox) with ESMTP id EA6F221731; Thu, 19 Nov 2026 20:20:03 +0100 (CET) From: Thomas Lamprecht To: pve-devel@lists.proxmox.com Subject: [PATCH cluster/guest-common/qemu-server/container/manager/docs v5 00/41] fix #7846: guest identity (uid) and cluster ID Date: Thu, 19 Nov 2026 20:18:32 +0100 Message-ID: <20261119191914.2336934-1-t.lamprecht@proxmox.com> X-Mailer: git-send-email 2.47.3 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Bm-Milter-Handled: 55990f41-d878-4baa-be0a-ee34c49e34d2 X-Bm-Transport-Timestamp: 1795115959046 X-SPAM-LEVEL: Spam detection results: 0 AWL 0.157 Adjusted score from AWL reputation of From: address DMARC_MISSING 0.1 Missing DMARC policy KAM_DMARC_STATUS 0.01 Test Rule for DKIM or SPF Failure with Strict Alignment (newer systems) RCVD_IN_DNSWL_MED -2.3 Sender listed at https://www.dnswl.org/, medium trust SPF_HELO_NONE 0.001 SPF: HELO does not publish an SPF Record SPF_PASS -0.001 SPF: sender matches SPF record Message-ID-Hash: TVD5I4AP4IC54VOR6RBX6SVE2PVVC6OS X-Message-ID-Hash: TVD5I4AP4IC54VOR6RBX6SVE2PVVC6OS X-MailFrom: t.lamprecht@proxmox.com X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header X-Mailman-Version: 3.3.10 Precedence: list List-Id: Proxmox VE development discussion List-Help: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: A VMID can be reused for another guest, and a guest can get another VMID or move to another cluster through a restore or a remote migration. So backup software and other tools cannot tell whether they still look at the same guest, see [0]. This series gives every VM and container a guest identity in the read-only 'meta' property: a random uid (100 bits as 20 base32 characters) and the time the guest was registered in the cluster. Create and clone get a new uid. Config changes, migration, rollback, backup and restore keep it, also under another VMID or in another cluster, as long as no other guest there has it. Each cluster or standalone node also gets a cluster ID in the same format, generated and protected by pmxcfs. The docs patches describe the details. The uid is a plain random nonce: 100 bits make a duplicate practically impossible and fit into a 128-bit value with room to spare, and 20 base32 characters stay short and need no quoting. The registration time is separate from the creation time, which clones, restores and remote migrations copy along. It changes when a guest gets registered again, so tools can tell such a guest from the one they knew. Three behavior changes come with it: * Restoring over an existing guest whose uid differs from the one of the backup is refused unless 'allow-identity-mismatch' is set. 'force' alone does not override this. * A migration to another cluster is refused if another guest there already has the uid, unless 'allow-new-uid' is set. The check runs before anything is transferred. PDM does not know the option yet. * Older versions drop the whole 'meta' property of a guest that has an identity. So migration and clone to such nodes is refused, and remote migration bumps the tunnel version. This also hits moves by the HA stack, which stops a container or a relocated VM before the migration gets refused. Making the HA stack aware of the node versions, so that for example the maintenance mode does not pick such nodes, is a separate follow-up. The patches are in dependency order: guest-common needs the pve-cluster of this series, qemu-server and container need guest-common, and manager needs all of them. manager also has to depend on the qemu-server and container versions with this series, as the migration guard takes the pve-manager version of a node as proof that the node keeps the identity. MIN_PVE_MANAGER_VERSION in guest-common is a stand-in for that version. Tested with the affected tests of the repositories for every patch and their full suites at the end, pmxcfs with corosync in network namespaces, and end to end on a nested three-node cluster that was upgraded node by node, with remote migrations to a standalone node. changes v4 -> v5, thanks again to Dominik for the review and the tests: * no lock of its own for the guest identity anymore, as Dominik noted too. A config write reads the stored identity and writes the config under the config lock of its caller, like every config change (patches 9, 10, 11). The lock was there because the container post-stop hook wrote the config without the config lock. That is a bug of the hook and fixed by an independent patch before the series, as is the same for the disk of a template in 'qm importdisk'. This was actually quite nasty pre-existing hole that was found here, nice. * remote migration from a cluster without guest identities: the task on the target logs the new uid of the guest, and that a migration back to such a cluster gets refused (patches 12, 18, 26) * new test for the cluster-wide bulk migration check (patch 34) * 'adopt' is the one word for getting the uid of a backup or of the source guest now, also in option descriptions, docs and subjects, where it was 'take over' before; the task logs already said 'adopt' * docs: the section about failed restores is a list of what can be left behind and numbered steps to clean up now (patch 36) * clearer wording of some comments, option descriptions and commit messages I added Dominik's Reviewed-by and Tested-by from v4 on all patches but 9, 10, 11, whose code changed with the removal of the lock, and 36, whose text got rewritten. changes v3 -> v4, thanks a lot (!) to Dominik for the review and the tests: * cluster ID: pmxcfs checks after each sync and then every ten seconds while in sync, by the current state alone, instead of on single events, which missed failed sends (the confdb listener patch is gone); a standalone node that cannot store its ID starts without one * failed restores: independent fixes before the series put the previous firewall config back, keep a guest whose cleanup was incomplete at its VMID and locked, for VMs as for containers, and do not create a guest for a caller that may only overwrite one * shared code: pve-guest-common has one helper for the start of a restore, both overrides as standard options and the wait for a volume transfer in the tunnel cleanup * tests of existing behavior moved to independent commits before the series, so both remote migration test patches left it * new, suggested by Dominik: the uid as variable of the backup notes template (patches 13, 35, 41) and the cluster ID in the datacenter summary (patch 30) * cluster-wide bulk migration between two older nodes works again, and the stand-in for the minimum pve-manager version is 9.2.33, as 9.2.31 and 9.2.32 got released without this series I added Dominik's Reviewed-by and Tested-by from v3 on the patches that did not change or only got polish since. changes v2 -> v3: * remote migration: a guest without uid gets one before the source sends it, so source and target no longer end up with unrelated uids; the identity command moved to pve-guest-common for both guest types (new patch 13) * restore: keep the guest if a restored volume or firewall config cannot be removed, fail if a container backup changes between its check and its extraction, and let the cleanup follow the reservation * reserving a VMID refuses an existing config that fails to load instead of replacing it * cluster ID: generate it right away when the configured nodes change (new patch 5) * feature API without version lookups for snapshot requests, and the cluster-wide bulk migration refuses its target up front * docs: failed restores, uid conflicts on migration, qm.conf(5) and pct.conf(5) * more independent fixes that the series builds on were applied on their own: failed and concurrent restores and creates, the schema of the migration preconditions, the lock of a VMA restore, a node list cut short in pmxcfs changes v1 -> v2, thanks to Dominik for the review: * rebase onto current master (tracking of used guest IDs, dfsm rework) * remote migration takes the uid over at its start and refuses a uid that is in use in the target cluster, unless 'allow-new-uid' is set * qemu-server and container share the restore checks and messages through pve-guest-common, with fewer special cases * IDs take each character from one random byte, without an encoder * the clone dialog does not offer nodes that would drop the identity, and a bulk migration checks its target once (patch 35) * independent fixes moved out of the series and got applied on their own: the IPC client error, the order of removing firewall and guest config, and the cleanup of a remote migration target after failures * comments, commit messages and docs give the reasons for the format, for rtime next to ctime and for the separate lock Details are in the notes below the "---" of each patch. v4: https://lore.proxmox.com/pve-devel/20261119103551.1666964-1-t.lamprecht@proxmox.com/ v3: https://lore.proxmox.com/pve-devel/20261118043817.1659212-1-t.lamprecht@proxmox.com/ v2: https://lore.proxmox.com/pve-devel/20261117223945.862914-1-t.lamprecht@proxmox.com/ v1: https://lore.proxmox.com/pve-devel/20261116004703.2908294-1-t.lamprecht@proxmox.com/ [0]: https://bugzilla.proxmox.com/show_bug.cgi?id=7846 cluster: Thomas Lamprecht (7): pmxcfs: reject changes to the cluster ID file from user space pmxcfs: generate a cluster ID on standalone nodes pmxcfs: dfsm: add a callback that runs once all members are in sync pmxcfs: quorum: record the corosync membership pmxcfs: dcdb: generate the cluster ID once the cluster is in sync cluster: add helper to read the cluster ID pmxcfs: allow removing the cluster ID file in forced local mode src/PVE/Cluster.pm | 16 + src/pmxcfs/Makefile | 12 +- src/pmxcfs/cfs-plug-memdb.c | 46 ++ src/pmxcfs/cfs-utils.c | 8 + src/pmxcfs/cfs-utils.h | 5 + src/pmxcfs/check_dcdb.c | 876 ++++++++++++++++++++++++++++++++++++ src/pmxcfs/check_memdb.c | 312 +++++++++++++ src/pmxcfs/cluster-id.c | 142 ++++++ src/pmxcfs/cluster-id.h | 50 ++ src/pmxcfs/dcdb.c | 141 ++++++ src/pmxcfs/dcdb.h | 5 +- src/pmxcfs/dfsm.c | 70 ++- src/pmxcfs/dfsm.h | 10 + src/pmxcfs/memdb.c | 17 + src/pmxcfs/memdb.h | 4 + src/pmxcfs/pmxcfs.c | 8 +- src/pmxcfs/quorum.c | 13 +- src/pmxcfs/status.c | 44 ++ src/pmxcfs/status.h | 13 + src/test/Makefile | 6 +- src/test/test_cluster_id.pl | 46 ++ 21 files changed, 1832 insertions(+), 12 deletions(-) create mode 100644 src/pmxcfs/check_dcdb.c create mode 100644 src/pmxcfs/cluster-id.c create mode 100644 src/pmxcfs/cluster-id.h create mode 100644 src/test/test_cluster_id.pl guest-common: Thomas Lamprecht (6): guest identity: add helpers for the guest uid and registration time abstract config: assign a guest identity and keep it on config writes abstract config: let the VMID reservation return the previous state guest identity: allow adopting a uid that is unused in the cluster migrate: add shared helpers for the identity command of the migration tunnel vzdump: notes template: list the uid variable src/Makefile | 1 + src/PVE/AbstractConfig.pm | 166 +++- src/PVE/AbstractMigrate.pm | 42 + src/PVE/GuestIdentity.pm | 514 +++++++++++ src/PVE/VZDump/Common.pm | 6 +- src/tests/Makefile | 10 +- src/tests/config-identity-tests.pl | 1341 ++++++++++++++++++++++++++++ src/tests/guest-identity-tests.pl | 474 ++++++++++ 8 files changed, 2545 insertions(+), 9 deletions(-) create mode 100644 src/PVE/GuestIdentity.pm create mode 100755 src/tests/config-identity-tests.pl create mode 100755 src/tests/guest-identity-tests.pl qemu-server: Thomas Lamprecht (6): api: restore: release the VMID if a restore to an unused VMID fails api: migrate, clone: refuse targets that would drop the guest identity fix #7846: config: add a guest identity to the 'meta' property restore: adopt the guest uid from the backup where possible api: remote migration: adopt the guest uid from the source api: status: return the guest uid src/PVE/API2/Qemu.pm | 108 ++- src/PVE/CLI/qm.pm | 1 + src/PVE/CLI/qmrestore.pm | 5 +- src/PVE/QemuConfig.pm | 7 + src/PVE/QemuMigrate.pm | 18 +- src/PVE/QemuServer.pm | 166 +++- src/PVE/QemuServer/MetaInfo.pm | 2 + src/test/Makefile | 5 +- src/test/QemuServerTest/FakeCFS.pm | 84 +- .../q35-windows-pinning-identity-only.conf | 6 + ...q35-windows-pinning-identity-only.conf.cmd | 26 + .../meta-identity-invalid.conf | 6 + .../meta-identity-invalid.conf.strict.error | 2 + .../meta-identity-invalid.conf | 7 + .../parse-config-input/meta-identity.conf | 20 + src/test/restore-config-expected/150.conf | 9 + src/test/restore-config-input/150.conf | 10 + src/test/run_api_tests.pl | 198 +++++ src/test/run_guest_identity_tests.pl | 630 +++++++++++++++ src/test/run_parse_config_tests.pl | 2 +- src/test/run_qemu_restore_config_tests.pl | 754 +++++++++++++++++- .../create/qemu-server/401.conf | 31 + .../create/qemu-server/402.conf | 31 + .../rollback/qemu-server/401.conf | 36 + .../rollback/qemu-server/402.conf | 35 + .../create/qemu-server/401.conf | 13 + .../create/qemu-server/402.conf | 13 + .../rollback/qemu-server/401.conf | 35 + .../rollback/qemu-server/402.conf | 34 + src/test/snapshot-test.pm | 82 +- 30 files changed, 2300 insertions(+), 76 deletions(-) create mode 100644 src/test/cfg2cmd/q35-windows-pinning-identity-only.conf create mode 100644 src/test/cfg2cmd/q35-windows-pinning-identity-only.conf.cmd create mode 100644 src/test/parse-config-expected/meta-identity-invalid.conf create mode 100644 src/test/parse-config-expected/meta-identity-invalid.conf.strict.error create mode 100644 src/test/parse-config-input/meta-identity-invalid.conf create mode 100644 src/test/parse-config-input/meta-identity.conf create mode 100644 src/test/restore-config-expected/150.conf create mode 100644 src/test/restore-config-input/150.conf create mode 100755 src/test/run_guest_identity_tests.pl create mode 100644 src/test/snapshot-expected/create/qemu-server/401.conf create mode 100644 src/test/snapshot-expected/create/qemu-server/402.conf create mode 100644 src/test/snapshot-expected/rollback/qemu-server/401.conf create mode 100644 src/test/snapshot-expected/rollback/qemu-server/402.conf create mode 100644 src/test/snapshot-input/create/qemu-server/401.conf create mode 100644 src/test/snapshot-input/create/qemu-server/402.conf create mode 100644 src/test/snapshot-input/rollback/qemu-server/401.conf create mode 100644 src/test/snapshot-input/rollback/qemu-server/402.conf container: Thomas Lamprecht (8): restore: decide by the reservation whether the container is new config: add read-only meta property with the creation time fix #7846: config: add a guest identity to the 'meta' property api: migrate, clone: refuse targets that would drop the guest identity restore: adopt the guest uid from the backup where possible restore: refuse overwriting a container with another guest's backup api: remote migration: adopt the guest uid from the source api: status: return the guest uid src/PVE/API2/LXC.pm | 136 +++- src/PVE/API2/LXC/Config.pm | 8 + src/PVE/API2/LXC/Status.pm | 6 + src/PVE/CLI/pct.pm | 2 + src/PVE/LXC/Config.pm | 42 +- src/PVE/LXC/Create.pm | 26 +- src/PVE/LXC/Migrate.pm | 8 +- src/test/config_tests.pl | 26 +- src/test/lifecycle_tests.pl | 1479 +++++++++++++++++++++++++++++++++++ 9 files changed, 1702 insertions(+), 31 deletions(-) manager: Thomas Lamprecht (8): api: cluster resources: add guest uid and registration time api: cluster status: add cluster ID ui: datacenter summary: show the cluster ID ui: restore: allow overwriting a guest with a backup of another guest ui: migrate: refuse target nodes that would drop the guest identity ui: guest summary: show the guest identity api: bulk migrate: refuse targets that would drop the guest identity vzdump: notes template: add the guest uid as variable PVE/API2/Cluster.pm | 24 +++- PVE/API2/Cluster/BulkAction/Guest.pm | 17 +++ PVE/API2/Nodes.pm | 3 + PVE/VZDump.pm | 12 +- test/BulkMigrate_test.pl | 102 ++++++++++++++ test/ClusterResources_test.pl | 171 +++++++++++++++++++++++ test/ClusterStatus_test.pl | 114 +++++++++++++++ test/Makefile | 14 +- www/manager6/Makefile | 1 + www/manager6/Utils.js | 2 +- www/manager6/dc/Health.js | 6 + www/manager6/panel/GuestIdentity.js | 120 ++++++++++++++++ www/manager6/panel/GuestStatusView.js | 3 + www/manager6/panel/TemplateStatusView.js | 3 + www/manager6/window/Migrate.js | 29 ++++ www/manager6/window/Restore.js | 18 +++ 16 files changed, 635 insertions(+), 4 deletions(-) create mode 100755 test/BulkMigrate_test.pl create mode 100755 test/ClusterResources_test.pl create mode 100755 test/ClusterStatus_test.pl create mode 100644 www/manager6/panel/GuestIdentity.js docs: Thomas Lamprecht (6): vzdump: restore: describe what a failed restore can leave behind vzdump: describe the guest identity of VMs and containers pvecm: document the cluster ID and how to get a new one qm: configuration: describe the read-only meta property pct: configuration: describe the read-only meta property vzdump: notes: document the uid template variable pct.adoc | 35 +++++ pct.conf.adoc | 3 + pmxcfs.adoc | 1 + pvecm.adoc | 106 +++++++++++++++ qm.adoc | 38 ++++++ qm.conf.adoc | 3 + vzdump.adoc | 354 ++++++++++++++++++++++++++++++++++++++++++++++++++ 7 files changed, 540 insertions(+) Summary over all repositories: 91 files changed, 9554 insertions(+), 132 deletions(-) -- 2.47.3