Bug#992213: release-notes: Release notes should mention lxc issues with cgroups change

[Mike Hommey]

Package: release-notes
Severity: important

The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Mike

[Paul Gevers]

Hi Mike,

Sorry for brevity, I'm in a hurry.

Post by Mike Hommey
The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Can you elaborate what the issues are you experience with LXC?

Paul

[Paul Gevers]

Hi Mike,

Post by Paul Gevers
Hi Mike,
Sorry for brevity, I'm in a hurry.

Post by Mike Hommey
The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Can you elaborate what the issues are you experience with LXC?

ping.

Paul

[Mike Hommey]

Post by Paul Gevers
Hi Mike,

Post by Paul Gevers
Hi Mike,
Sorry for brevity, I'm in a hurry.

Post by Mike Hommey
The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Can you elaborate what the issues are you experience with LXC?

ping.

Sorry, I somehow missed your email.

Containers would just not start, with messages like:
Failed to mount cgroup at /sys/fs/cgroup/systemd: Operation not permitted
[!!!!!!] Failed to mount API filesystems.
Exiting PID 1...

Mike

[Paul Gevers]

Hi Mike,

Post by Mike Hommey

Post by Paul Gevers
Hi Mike,

Post by Paul Gevers
Hi Mike,
Sorry for brevity, I'm in a hurry.

Post by Mike Hommey
The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Can you elaborate what the issues are you experience with LXC?

ping.

Sorry, I somehow missed your email.
Failed to mount cgroup at /sys/fs/cgroup/systemd: Operation not permitted
[!!!!!!] Failed to mount API filesystems.
Exiting PID 1...

Interesting. We heavily use lxc on ci.d.n (all tests run in one) and we
haven't experienced this. I'm wondering what the specifics of you
system(s) is that trigger the issue. Do you have ideas?

I'm wondering if it's not related to the discussion we had in bug
985617: linux-user-namespaces. See
https://www.debian.org/releases/bullseye/amd64/release-notes/ch-information.en.html#linux-user-namespaces

Paul

[Mike Hommey]

Post by Paul Gevers
Hi Mike,

Post by Mike Hommey

Post by Paul Gevers
Hi Mike,

Post by Paul Gevers
Hi Mike,
Sorry for brevity, I'm in a hurry.

Post by Mike Hommey
The release notes has a section about the issues with openstack, but
there are also problems with lxc. I'm not sure what the proper
workaround is, but setting `lxc.mount.auto = cgroup:mixed:force` fixed
it for me.

Can you elaborate what the issues are you experience with LXC?

ping.

Sorry, I somehow missed your email.
Failed to mount cgroup at /sys/fs/cgroup/systemd: Operation not permitted
[!!!!!!] Failed to mount API filesystems.
Exiting PID 1...

Interesting. We heavily use lxc on ci.d.n (all tests run in one) and we
haven't experienced this. I'm wondering what the specifics of you
system(s) is that trigger the issue. Do you have ideas?

I actually don't. I have some containers that trigger it, and some that
don't. I wasn't able to figure out why.

Mike

[Paul Gevers]

Hi Mike,

Post by Mike Hommey

Post by Paul Gevers
Interesting. We heavily use lxc on ci.d.n (all tests run in one) and we
haven't experienced this. I'm wondering what the specifics of you
system(s) is that trigger the issue. Do you have ideas?

I actually don't. I have some containers that trigger it, and some that
don't. I wasn't able to figure out why.

That doesn't make an awful good release-notes report. Have you filed the
bug somewhere (e.g. against the lxc package). I think that's probably
the better course of action than to try to figure it out with the
release notes editors.

Paul

[Paul Gevers]

Hi Mike,

Post by Paul Gevers

Post by Mike Hommey

Post by Paul Gevers
Interesting. We heavily use lxc on ci.d.n (all tests run in one) and we
haven't experienced this. I'm wondering what the specifics of you
system(s) is that trigger the issue. Do you have ideas?

I actually don't. I have some containers that trigger it, and some that
don't. I wasn't able to figure out why.

That doesn't make an awful good release-notes report. Have you filed the
bug somewhere (e.g. against the lxc package). I think that's probably
the better course of action than to try to figure it out with the
release notes editors.

Did you do anything with this? Without more information, this bug isn't
really actionable.

Paul

[Debian Bug Tracking System]

Your message dated Sat, 4 Mar 2023 20:44:20 +0100
with message-id <916d165a-5cd2-121c-72f0-***@debian.org>
and subject line Re: Bug#992213: release-notes: Release notes should mention lxc issues with cgroups change
has caused the Debian Bug report #992213,
regarding release-notes: Release notes should mention lxc issues with cgroups change
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ***@bugs.debian.org
immediately.)