During preparation for rollback, there already is a VM cleanup. The
rollback operation temporarily drops the file lock for the
configuration while doing rollback of the volumes on the storage
layer. If a shutdown operation or more precisely, the cleanup
thereafter happens concurrently, it can get the file lock and do the
cleanup (a second time) while rollback is running. If a badly timed
deactivate for the volume happens, this can cause issues as reported
in the community forum [0] and confirmed locally (but only managed to
reproduce with 'saferemove').

[0]: https://forum.proxmox.com/threads/174875/post-815273

Signed-off-by: Fiona Ebner <[email protected]>
---
 src/PVE/CLI/qm.pm | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/src/PVE/CLI/qm.pm b/src/PVE/CLI/qm.pm
index bfce860f..60fe318e 100755
--- a/src/PVE/CLI/qm.pm
+++ b/src/PVE/CLI/qm.pm
@@ -1111,6 +1111,12 @@ __PACKAGE__->register_method({
                 my $pid = PVE::QemuServer::check_running($vmid);
                 die "vm still running\n" if $pid;
 
+                # Rollback already does cleanup when preparing and afterwards 
temporarily drops the
+                # lock on the configuration file to rollback the volumes. 
Deactivating volumes here
+                # again while that is happening would be problematic.
+                die "skipping cleanup - 'rollback' lock is present\n"
+                    if $conf->{lock} && $conf->{lock} eq 'rollback';
+
                 if (!$clean) {
                     # we have to cleanup the tap devices after a crash
 
-- 
2.47.3



_______________________________________________
pve-devel mailing list
[email protected]
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Reply via email to