[pve-devel] [PATCH container/qemu-server v6 0/7] fix #3711 and adapt drive detach/remove behavior

From: "Michael Köppl" <m.koeppl@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [pve-devel] [PATCH container/qemu-server v6 0/7] fix #3711 and adapt drive detach/remove behavior
Date: Tue, 20 May 2025 11:08:11 +0200	[thread overview]
Message-ID: <20250520090818.44881-1-m.koeppl@proxmox.com> (raw)

This series aims to fix #3711 [0] and streamline the detach/remove
behavior around volumes that are either mounted into a container or
attached to a VM as a hard disk. It is a continuation of a series from
2022 [1], but makes the following changes:

Changes v5 -> v6:
- Fix links in cover letter
- Use Originally-by instead of Co-authored-by
- Add documentation for the second patch regarding the use of
  $mp->{volume} instead of $conf->{$opt}
- Note that no functional changes are intended for the the second patch

Changes v4 -> v5:
- Always ignore errors that originate from a removed storage and
  continue with destruction of a container or removal of a volume, 
  instead of adding an option to ignore these errors.
- Remove web UI checkbox
- Remove formatting patch
- Additionally allow removing a mount point with a removed storage
  from a running container (previously hotplug removal was not possible)
- Fix style nits from v4 review
- Print warnings for any errors that occur instead of ignoring them
- Add explicit check if storage still exists when destroying a container
  to differentiate between that case and other error cases (which should
  still fail)

The first change also means that this patch series should be held back
until PVE 9. In the original patch series the option to ignore these
errors was made to avoid breaking existing behavior. After some off-list
discussion it seems more reasonable to avoid an additional option and
instead allow users to remove containers with mount points even if the
underlying storage of the mount point was removed.

Where at least some of the implementation was taken from the previous
series, the patch was marked to be Originally-by by the original author.

[0] https://bugzilla.proxmox.com/show_bug.cgi?id=3711
[1] https://lore.proxmox.com/pve-devel/20221125144008.2988072-1-s.hrdlicka@proxmox.com/t/#

pve-container:

Michael Köppl (4):
  fix #3711: lxc: print warning if storage for mounted volume does not
    exist anymore
  config: apply_pending: get unused volid through parse_volume()
  fix #3711: lxc: allow removing unused mp if storage no longer exists
  add linked clone check when destroying container

 src/PVE/LXC.pm        | 24 +++++++++++++++++++++++-
 src/PVE/LXC/Config.pm | 25 ++++++++++++++++++++-----
 2 files changed, 43 insertions(+), 6 deletions(-)

qemu-server:

Michael Köppl (3):
  adapt linked clone check to not die if an error occurs during check
  print warning for PVE::Storage::path errors instead of failing
  mark volumes pending detach as unused if storage was removed

 PVE/QemuServer.pm | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

Summary over all repositories:
  3 files changed, 54 insertions(+), 11 deletions(-)

-- 
Generated by git-murpp 0.8.0

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