Closed medyagh closed 3 years ago
This looks like the exact same issue that we were seeing with podman in #8033
The real error isn't actually the re-attaching to the old cgroups (since those are gone). We see this all the time, even when it starts up correctly again - for both drivers.
The problem is about failing to create some fiiles, possibly due to read-only /sys
.
These lines are supposed to be enough:
INFO: remounting /sys read-only
INFO: fix cgroup mounts for all subsystems
And there doesn't seem to be any errors ?
That should make /sys/fs/cgroup/systemd
writable, for the docker/<container>
subdirectory.
But here it is still failing, the second time ?
2020-05-09T01:13:33.3937272Z �[0;1;31mFailed to open pin file: No such file or directory�[0m
2020-05-09T01:13:33.3937576Z �[0;1;31mFailed to allocate manager object: No such file or directory�[0m
2020-05-09T01:13:33.3937846Z [�[0;1;31m!!!!!!�[0m] Failed to allocate manager object.
2020-05-09T01:13:33.3938097Z �[0;1;31mExiting PID 1...�[0m
For podman the issue was just that KIND only did the workaround for "docker", not "libpod_parent"
For debugging this, you can get the ID of the minikube container and then look at the mount:
findmnt /sys/fs/cgroup/systemd/docker/${CONTAINER_ID}
It supposed to be mounted as cgroup rw
, even though systemd requires /sys
to be ro
.
Seems like fix_cgroups
(in entrypoint) silently failed on us ?
Doesn't seem to reproduce locally, can stop/start docker containers multiple times
So it must be something specific to Azure Linux, or maybe just a fluke perhaps...
Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale
.
Stale issues rot after an additional 30d of inactivity and eventually close.
If this issue is safe to close now please do so with /close
.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta. /lifecycle stale
Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten
.
Rotten issues close after an additional 30d of inactivity.
If this issue is safe to close now please do so with /close
.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta. /lifecycle rotten
this happens only in github actions when trying to start a stopped kic. as seen in this PR: https://github.com/kubernetes/minikube/pull/8025/checks?check_run_id=658120043