OOM killer activated after internal move Rocky VM

Rocky Linux Help & Support
1

When moving some VMs between hosts (XCP-ng virtualization), the OOM (Out-Of-Memory) killer is triggered, and the VM no longer behaves correctly afterwards.

oomkiller
oomkiller771×404 8.36 KB

In this particular case, it is our Git-ICT server. At first sight, the system appears to be working normally, but some functionality is no longer behaving as expected.

We have to reboot the system afterwards in order to restore normal operation.

Does anybody have the same problem?

2

Does the VM have the XCP-ng guest agent installed? Sounds like an XCP problem which is causing the OOM to happen.

3

Is this log from the guest or the host?
From the screenshot, I don’t see the oom killer “triggered”, I mean it becomes enabled and disabled, but I don’t see it has killed any pid? The oom killer is usually only an issue if you overcommit memory, check the policy.