random kernel panics

[Tim Korn]

I have two hosts that experience kernel panics intermittently.

I have not been able to find any pattern. The occurrences seem random across both hosts, and the console logs show multiple different processes associated.

The pertinent kernel logs from the most recent panics are included in the attachment.

These hosts are openvpn servers which have 500+ users connected. I have two additional servers (essentially identical configuration) with far fewer users connected and these have never experienced a a kernel panic.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-6.8.0-1033-gcp 6.8.0-1033.35~22.04.1
ProcVersionSignature: Ubuntu 6.8.0-1033.35~22.04.1-gcp 6.8.12
Uname: Linux 6.8.0-1033-gcp x86_64
ApportVersion: 2.20.11-0ubuntu82.9
Architecture: amd64
CasperMD5CheckResult: unknown
CloudArchitecture: x86_64
CloudID: gce
CloudName: gce
CloudPlatform: gce
CloudRegion: us-central1
CloudSubPlatform: metadata (http://metadata.google.internal/computeMetadata/v1/)
Date: Tue Jan 6 17:27:51 2026
ProcEnviron:
 SHELL=/bin/bash
 LANG=C.UTF-8
 TERM=xterm-256color
 PATH=(custom, no user)
SourcePackage: linux-signed-gcp-6.8
UpgradeStatus: No upgrade log present (probably fresh install)

[Tim Korn (kornface13)]

Found this is related to inode switching. THis is the second time we're run into this kernel issue on different systems.

$ sudo bpftrace test.bt
Attaching 1 probe...
[18:27:58] inode is switching! inode: 2921931 old cgroup: 92926276 new cgroup: 88
[18:27:59] inode is switching! inode: 2921933 old cgroup: 92926396 new cgroup: 88
[18:27:59] inode is switching! inode: 2921934 old cgroup: 92926498 new cgroup: 88
[18:27:59] inode is switching! inode: 2921936 old cgroup: 92926618 new cgroup: 88
[18:27:59] inode is switching! inode: 2921935 old cgroup: 92926558 new cgroup: 88
$ sudo find /root -inum 2921931
/root/.config/gcloud/logs/2026.01.06/18.27.52.779901.log
$ sudo find /root/.config/gcloud/logs/ | wc -l
2216568

The culprit was a bad script that caused excessive logging, however, this still shouldn't cause a kernel panic.