can't open '/sys/fs/cgroup/qemu.slice/2113.scope//memory.current' - No such file or directory (500)

RootS3c

New Member
Jan 2, 2026
8
0
1
After upgrading from qemu-server=9.0.30 i get this error on every VM

can't open '/sys/fs/cgroup/qemu.slice/2113.scope//memory.current' - No such file or directory (500)

cgroups2 works
 
the patch is already applied, so it will be included in the next bump for qemu-server


when you manually apply patches you have to reload/restart pveproxy & pvedaemon, so they load the perl libraries again
We encountered this during a restore of a VM with Veeam. All the other VMs became with status unknown in the GUI. qm list was failing as well

After applying the fix, the error is now reported only for that vmid, and the rest of the vms listed correctly in both GUI and qm list.

However we are seeing these new errors
Code:
Mar 05 11:37:37 tst01pvepoc01 pvestatd[62833]: qemu status update error: can't open '/sys/fs/cgroup/qemu.slice/122.scope//memory.current' - No such file or directory
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: unable to get memory stat for 122 - can't open '/sys/fs/cgroup/qemu.slice/122.scope//memory.current' - No such file or directory
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2599.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2600.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2601.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2602.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2603.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2604.
 
We encountered this during a restore of a VM with Veeam. All the other VMs became with status unknown in the GUI. qm list was failing as well

After applying the fix, the error is now reported only for that vmid, and the rest of the vms listed correctly in both GUI and qm list.

However we are seeing these new errors
Code:
Mar 05 11:37:37 tst01pvepoc01 pvestatd[62833]: qemu status update error: can't open '/sys/fs/cgroup/qemu.slice/122.scope//memory.current' - No such file or directory
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: unable to get memory stat for 122 - can't open '/sys/fs/cgroup/qemu.slice/122.scope//memory.current' - No such file or directory
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2599.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2600.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2601.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2602.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2603.
Mar 05 11:37:45 tst01pvepoc01 pvedaemon[124152]: Use of uninitialized value in multiplication (*) at /usr/share/perl5/PVE/QemuServer.pm line 2604.

We are having the exact same issue restoring from Veeam. Running Proxmox 9.1.6 with the enterprise repository.

Funny thing is this error occurs only on our Dell R7525 hosts with AMD CPUs, and not on our dev servers which are Dell R730 hosts which have Intel CPUs in them.

To work around it I'm having to restore to our dev environment and then I am using Proxmox Datacenter Manager to migrate from the dev cluster to the production cluster.
 
We're seeing exactly the same thing when restoring a VMware vm to proxmox using Veeam. Server is a Dell Poweredge R740 with intel CPU's.
After the attempted restore all VM's in the list show with a ?
Veeam restore fails and the restored VM wont start with the same error.

1772893694429.png

1772893865979.png
 
Proxmox 9.1.4–9.1.6 shows memory.current errors after Veeam restore. Patch exists, but you need to restart pveproxy and pvedaemon after applying. Workarounds: downgrade to 9.1.3, restore on a dev cluster then migrate, or wait for the next qemu-server update. Mostly affects AMD hosts.
Yes - I resolved display issues by patching the perl in QemuServer.pm.
Restore still failed with Veeam backups vmware > Prox however. The only way i could get this to work was to upgrade Veaam to version 13 (we were on 12) after which everything worked as it should.
 
after spending a few hours trying to track this down i found this thread - i was able to confirm that qemu-server 9.1.3 works and 9.1.4 is the version that causes the crash when doing Veeam restores


for anyone else looking for the exact command - i was not sure how to apply the patch in the link above, i dont see any fix in 9.1.5 test either
apt-get install qemu-server=9.1.3
 
Hi @gasherbrum,
did you actually test 9.1.5? The fix is in there:
Code:
qemu-server (9.1.5) trixie; urgency=medium
  ...
  * vm status: guard memory stat collection with eval to avoid breaking the
    whole VM status API call or pvestatd collection if it fails.
 
@fiona confirmed that 9.1.5 is working correctly with Veeam, there are the erroneous errors present in rborgs post but those also exist on 9.1.3, thank you
 
Hi @fiona and @gasherbrum - would you help a newbie? When I attempt

apt-get install qemu-server=9.1.5

I get "Version '9.1.5' for 'qemu-server' was not found. I can find older versions however.

Thanks so much!
-tj
you have to enable the test repository for 9.1.5, i think its not available yet on no-subscription