Backport the patches for this issue:
https://bugzilla.redhat.com/show_bug.cgi?id=1968528
It reports:
The fix for Bug 1819868 has introduced a new issue related to its
implementation of rate limiting.
Rate limiting the mount_event_source can cause unmount events to be
missed, which leads to mount unit cgroups being leaked (not cleaned up
when the mount is gone).
The fix for 1968528 can fix the issue we met:
During the reboot process of subclouds (either lock-unlock or sudo
reboot), unmounting failure messages repeat for a few hundred of times.
The patches are listed at:
https://github.com/redhat-plumbers/systemd-rhel8/pull/198/commits
And they are picked from https://github.com/systemd-rhel/rhel-8/ (branch
rhel-8.4.0).
Verification:
In my test on an AIO-SX lab, the bug appears as:
run "sudo reboot" on controller, endless unmounting failure logs
printed.
Verified that the problem was there during the shutdown
phase of a reboot. Reinstalled with a fixed image, and verified that
the issue was now gone by doing 5 reboots. Ran sanity on the lab,
and verified no new issues seen.
Closes-Bug: #1948899
Signed-off-by: Li Zhou <li.zhou@windriver.com>
Change-Id: If95932ceead1bea973f2219d3a8d6b04cf0fd5f8
bed1e46362