"Resume" action fails with Services not running that should be: ceilometer-agent-compute' when upgrading one-by-one from Ussuri to Victoria

Bug #2067492 reported by Aliaksandr Vasiuk
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack Nova Compute Charm
Invalid
Undecided
Unassigned

Bug Description

Hi,

We noticed today, that when upgrading a cloud from Ussuri to Victoria in "one by one" manner, `resume` action fails with:
```
message: 'Action resume failed: Couldn''t resume: ceilometer-agent-compute didn''t
  resume cleanly.; Services not running that should be: ceilometer-agent-compute'
```

So, the cloud is:
* charmed openstack Focal Ussuri in a middle of upgrade to Focal Victoria.
* Control Plane is all Victoria.
* Juju 2.9.45
* nova-compute charm is: victoria/stable 727
* apt nova-compute packages are: 2:21.2.4-0ubuntu2.6
* all machines are Focal 20.04.6

Steps to reproduce:
* Pause nova unit
```
juju run-action --wait nova-compute/2 pause
```
* Resume nova-unit
```
juju run-action --wait nova-compute/2 resume
unit-nova-compute-2:
  UnitId: nova-compute/2
  id: "5963"
  message: 'Action resume failed: Couldn''t resume: ceilometer-agent-compute didn''t
    resume cleanly.; Services not running that should be: ceilometer-agent-compute'
  status: failed
  timing:
    completed: 2024-05-29 14:06:03 +0000 UTC
    enqueued: 2024-05-29 14:05:52 +0000 UTC
    started: 2024-05-29 14:05:52 +0000 UTC
```
* Nova unit is left in Blocked state.

I was able to reproduce it on:
* All nova units before "juju run <> openstack-upgrade"
* All nova units that was paused on Ussuri, then got upgraded with "juju run <> openstack-upgrade", then resume fails when unit is on Victoria
But when I pause and resume a unit that was already upgraded then everything is fine.

In case you need it: full output of failed resume action: https://pastebin.ubuntu.com/p/wKkXnDB3sN/

Revision history for this message
Felipe Reyes (freyes) wrote : Re: [Bug 2067492] [NEW] "Resume" action fails with Services not running that should be: ceilometer-agent-compute' when upgrading one-by-one from Ussuri to Victoria

"""
    Failed to start ceilometer-agent-compute.service: Unit nova-compute.service is masked.
"""

I believe this is an ocurrence of this bug https://bugs.launchpad.net/charm-ceilometer-agent/+bug/1952882 , that's why newer releases of ceilometer-agent no longer have a pause/resume action - https://charmhub.io/ceilometer-agent/actions

Revision history for this message
Aliaksandr Vasiuk (valexby) wrote :
Changed in charm-nova-compute:
status: New → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.