[Bug 1944004] Re: snapd.seeded.service never finishes on non-amd64

[John Chittum]

I've added `livecd-rootfs` as affected after discussing the issue with
the snapd team. I'll be introducing a change to `livecd-rootfs` to
address how snapd expects cgroup2 with a mount at /sys/fs/cgroup

** Also affects: livecd-rootfs (Ubuntu)
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to livecd-rootfs in Ubuntu.
https://bugs.launchpad.net/bugs/1944004

Title:
  snapd.seeded.service never finishes on non-amd64

Status in cloud-images:
  Confirmed
Status in docker.io package in Ubuntu:
  Invalid
Status in livecd-rootfs package in Ubuntu:
  New
Status in snapd package in Ubuntu:
  Fix Released

Bug description:
  When launching an lxd container on non-amd64 architectures,
  snapd.seeded.service hangs during "boot" time and never finishes,
  which ends up impacting other systemd units.

  This bug was found while investigating what I thought was a cloud-init
  problem, because after launching a container I'd see:

  # cloud-init status
  status: running

  and the status would never transition to "done".  James Falcon (cloud-
  init dev) helped with the investigation and found that snapd was
  actually the culprit here.  One can check it by doing:

  # systemd-analyze blame
  Bootup is not yet finished (org.freedesktop.systemd1.Manager.FinishTimestampMonotonic=0).
  Please try again later.
  Hint: Use 'systemctl list-jobs' to see active jobs
  # systemctl list-jobs 
  JOB UNIT                                 TYPE  STATE  
  102 cloud-init.target                    start waiting
  110 cloud-final.service                  start waiting
  2   multi-user.target                    start waiting
  131 snapd.autoimport.service             start waiting
  103 cloud-config.service                 start waiting
  1   graphical.target                     start waiting
  111 snapd.seeded.service                 start running
  141 systemd-update-utmp-runlevel.service start waiting

  8 jobs listed.
  # systemctl status snapd.seeded.service 
  ● snapd.seeded.service - Wait until snapd is fully seeded
       Loaded: loaded (/lib/systemd/system/snapd.seeded.service; enabled; vendor preset: enabled)
       Active: activating (start) since Fri 2021-09-17 19:46:40 UTC; 24min ago
     Main PID: 336 (snap)
        Tasks: 8 (limit: 4777)
       Memory: 35.6M
       CGroup: /system.slice/snapd.seeded.service
               └─336 /usr/bin/snap wait system seed.loaded

  Sep 17 19:46:40 docker systemd[1]: Starting Wait until snapd is fully
  seeded...

  
  Inspecting snapd.service's logs, I see some strange messages (for example, a segmentation fault on taskrunner.go) that could be related to this problem.  I'm attaching the logs to this bug.

  This issue is blocking docker.io on impish-proposed, which in turn is
  impacting the build of our OCI images for Ubuntu 21.10.

To manage notifications about this bug go to:
https://bugs.launchpad.net/cloud-images/+bug/1944004/+subscriptions