Ceph-dashboard charm stuck waiting while status log says its blocked
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Invalid
|
Undecided
|
Unassigned | ||
Ceph Dashboard Charm |
Fix Released
|
High
|
Unassigned |
Bug Description
I'm not sure how to word a good description for this bug.
In test run https:/
Juju wait (I know, we need to switch to juju wait-for but since juju wait calls `juju status` this is still a valid concern), says that 2 ceph-dashboard units have been stuck waiting since `2023-08-24 06:38:02+00:00`.
Indeed, when we dump the `juju status` output (timestamped at `10:29:48Z`) it shows the ceph-dashboard units are waiting with a message "Charm configuration in progress".
In the juju-crashdump we collect the status logs, which shows the following:
====
bas@bas:
Time Type Status Message
24 Aug 2023 05:16:55Z juju-unit executing running loadbalancer-
24 Aug 2023 05:16:59Z juju-unit executing running loadbalancer-
24 Aug 2023 05:17:04Z juju-unit executing running loadbalancer-
24 Aug 2023 05:17:09Z juju-unit executing running loadbalancer-
24 Aug 2023 05:17:14Z juju-unit executing running loadbalancer-
24 Aug 2023 05:17:15Z juju-unit idle
24 Aug 2023 05:19:05Z juju-unit executing running loadbalancer-
24 Aug 2023 05:19:07Z juju-unit idle
24 Aug 2023 06:28:24Z juju-unit executing running certificates-
24 Aug 2023 06:28:25Z juju-unit idle
24 Aug 2023 06:29:19Z juju-unit executing running certificates-
24 Aug 2023 06:29:27Z juju-unit executing running dashboard-
24 Aug 2023 06:29:30Z juju-unit idle
24 Aug 2023 06:32:15Z juju-unit executing running radosgw-
24 Aug 2023 06:32:16Z juju-unit executing running radosgw-
24 Aug 2023 06:32:17Z juju-unit executing running radosgw-
24 Aug 2023 06:32:18Z juju-unit executing running radosgw-
24 Aug 2023 06:32:20Z juju-unit idle
24 Aug 2023 06:32:41Z workload blocked Dashboard is not enabled
24 Aug 2023 10:33:02Z workload waiting Charm configuration in progress
====
This says that the charm is not waiting, but has actually been blocked since 6:32 and would still be blocked when we dumped the juju status. According to this, it only switched to waiting one minute after we viewed the juju status.
This bug feels very similar to LP: #1992666. It may be an issue in the ceph-dashboard charm instead, but with the conflicting information from juju it is hard to open a bug there.
Extra logs and config files can be found here: https:/
tags: | added: cdo-qa foundations-engine |
Changed in charm-ceph-dashboard: | |
status: | Incomplete → In Progress |
importance: | Undecided → High |
assignee: | nobody → Alex Kavanagh (ajkavanagh) |
Can you provide a snippet of yaml status to show the application status data and the status data for each unit. It could be related to juju combining the unit statuses to get an overall app status if the app status is not explicitly set.