Re: [PATCH qemu-server v2] fix #7119: qm cleanup: wait for process exiting for up to 30 seconds

[Fiona Ebner <f.ebner@proxmox.com>]

Am 13.02.26 um 1:20 PM schrieb Fabian Grünbichler:
> On February 13, 2026 1:14 pm, Fiona Ebner wrote:
>> Am 10.02.26 um 12:14 PM schrieb Dominik Csapak:
>>> +                my $timeout = 30;
>>> +                my $starttime = time();
>>>                  my $pid = PVE::QemuServer::check_running($vmid);
>>> -                die "vm still running\n" if $pid;
>>> +                warn "vm still running - waiting up to $timeout seconds\n" if $pid;
>>
>> While we're at it, we could improve the message here. Something like
>> 'QEMU process $pid for VM $vmid still running (or newly started)'
>> Having the PID is nice info for developers/support engineers and the
>> case where a new instance is started before the cleanup was done is also
>> possible.
>>
>> In fact, the case with the new instance is easily triggered by 'stop'
>> mode backups. Maybe we should fix that up first before adding a timeout
>> here?
>>
>> Feb 13 13:09:48 pve9a1 qm[92975]: <root@pam> end task
>> UPID:pve9a1:00016B30:000CDF80:698F1485:qmshutdown:102:root@pam: OK
>> Feb 13 13:09:48 pve9a1 systemd[1]: Started 102.scope.
>> Feb 13 13:09:48 pve9a1 qmeventd[93079]: Starting cleanup for 102
>> Feb 13 13:09:48 pve9a1 qmeventd[93079]: trying to acquire lock...
>> Feb 13 13:09:48 pve9a1 vzdump[92895]: VM 102 started with PID 93116.
>> Feb 13 13:09:48 pve9a1 qmeventd[93079]:  OK
>> Feb 13 13:09:48 pve9a1 qmeventd[93079]: vm still running
> 
> does this mean we should actually have some sort of mechanism similar to
> the reboot flag to indicate a pending cleanup, and block/delay starts if
> it is still set?

Blocking/delaying starts is not what happens for the reboot flag/file:

> Feb 13 14:00:16 pve9a1 qm[124470]: <root@pam> starting task UPID:pve9a1:0001E639:001180FE:698F2060:qmreboot:102:root@pam:
> Feb 13 14:00:16 pve9a1 qm[124472]: <root@pam> starting task UPID:pve9a1:0001E63A:0011811E:698F2060:qmstart:102:root@pam:
> Feb 13 14:00:16 pve9a1 qm[124474]: start VM 102: UPID:pve9a1:0001E63A:0011811E:698F2060:qmstart:102:root@pam:
> [...]
> Feb 13 14:00:22 pve9a1 systemd[1]: 102.scope: Deactivated successfully.
> Feb 13 14:00:22 pve9a1 systemd[1]: 102.scope: Consumed 2min 3.333s CPU time, 2G memory peak.
> Feb 13 14:00:23 pve9a1 qmeventd[124565]: Starting cleanup for 102
> Feb 13 14:00:23 pve9a1 qmeventd[124565]: trying to acquire lock...
> Feb 13 14:00:23 pve9a1 qm[124470]: <root@pam> end task UPID:pve9a1:0001E639:001180FE:698F2060:qmreboot:102:root@pam: OK
> Feb 13 14:00:23 pve9a1 systemd[1]: Started 102.scope.
> Feb 13 14:00:23 pve9a1 qm[124474]: VM 102 started with PID 124620.
> Feb 13 14:00:23 pve9a1 qmeventd[124565]:  OK
> Feb 13 14:00:23 pve9a1 qmeventd[124565]: vm still running

Currently, it's just indicating whether the cleanup handler should start
the VM again afterwards.

Am 13.02.26 um 1:22 PM schrieb Dominik Csapak:
> Sounds good, one possibility would be to do no cleanup at all when doing
> a stop mode backup?
> We already know we'll need the resources (pid/socket/etc. files, vgpus,...) again?
> 
> Or is there some situation where that might not be the case? 

We do it for reboot (if not another start task sneaks in like in my
example above), and I don't see a good reason from the top of my head
why 'stop' mode backup should behave differently from a reboot (for
running VMs). It even applies pending changes just like a reboot right now.

I'm not sure if there is an actual need to do cleanup or if we could
also skip it when we are planning to spin up another instance right
away. But we do it for reboot, so the "safe" variant is also doing it
for 'stop' mode backup. History tells me it's been there since the
reboot functionality was added:
https://lists.proxmox.com/pipermail/pve-devel/2019-September/038988.html