Description of the problem: Cluster status in ACM shows message ( Upgrading to 4.X.Y (80%) while the real installed version is higher then Y (4.10.11 upgrading to 4.10.8 and 4.9.29 to 4.9.27, percent status may vary). More on a Screenshot-2022-05-09. Status of affected cluster in OCP seems fine, must-gather of affected cluster looks OK. Release version: ACM 2.4 Operator snapshot version: OCP version: Browser Info: Steps to reproduce: 1. 2. 3. Actual results: Expected results: Versions showed in ACM are in harm with managed cluster. Additional info: Customer wants to cooperate on finding root of problem. Recommended re-import as a workaround is not solution for him now.
When upgrading a managed cluster, cluster-curator-controller will create a job to monitor the progress of the upgrading and update the status of the clustercurator CR accordingly. If the job fails for some reason, the status of the clustercurator CR will stuck in a stale status, like `Upgrade status - Working towards 4.10.8: 620 of 771 done (80% complete)`, while the upgrading may have already been completed. ACM console reads the upgrading status from the clustercurator CR, that's the reason why cluster is showing a message that it is upgrading to a lower version. The issues has been fixed in 2.5 release, and the fix should be backport to 2.4 as well.
Thank you... This is helpful. And, please, do you know which 2.4 release will contain this fix?
The fix will be included in 2.4.6
If I understand well, the upgrade solve root of issue - the same thing will not happen again. But, what to do if I am facing this issue - please, how to solve it?
If the issue has already happened, you can just delete the clustercurator resource from the cluster namespace on the hub cluster with command below, oc -n <managed_cluster_name> delete clustercurator <managed_cluster_name>
The fix has been merged. It will be available in ACM 2.4.6.
G2Bsync 1281228723 comment Randy424 Mon, 17 Oct 2022 17:41:41 UTC G2Bsync Hey, it seems we missed the boat for 2.4.8 code freeze. We have some testing debt to cover before merging to the 2.4 branch, (the new sonar cloud coverage requires us to). Do we need this fix in 2.4.8? If the customer is still on this version (2.4.z), we could override our test threshold to get it in the release with your approval @cadawson. Otherwise, obviously we are looking at a 2.4.9 timeframe. Thanks. FYI @KevinFCormier (reposted for G2Bsync)
G2Bsync 1281237682 comment Randy424 Mon, 17 Oct 2022 17:48:46 UTC G2Bsync Also @cadawson, @KevinFCormier noticed that the original BZ tickets have been closed for this issue, maybe the need for a 2.4 backport is less dire than we think?
@leyan UI fixes are delivered, but I may have moved this to ON_QA prematurely - is there something for ClusterCurator you are still working on?
Hi @kcormier just wanted to confirm if 2.4.9 is the version containing the UI fixes
@nagbetra Hi, this fix was merged in to 2.4.9. See our github thread: https://github.com/stolostron/backlog/issues/22692#issuecomment-1307661310 Thanks Napoco!
Verified on 2.4.9-DOWNSTREAM-2022-11-17-20-19-31
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Red Hat Advanced Cluster Management 2.4.9 bug fixes and container updates), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2022:8656