Description of problem: I had to tear down my OCP 3 cluster, and when I brought back the new one and fixed the SA token using mig-ui, I wasn't able to get a good response. The only way I fixed this was by restarting mig-controller. Version-Release number of selected component (if applicable): Exists in 1.4.2 and in master. How reproducible: Seems to be consistent. Steps to Reproduce: 1. Set up a MigCluster with working SA token 2. Break the SA token (e.g. by removing the cluster, or just editing the token) 3. Try to fix the SA token in mig-ui or manually 4. See failure for MigCluster to become ready This may be a validation ordering problem: as seen below, the entire reconcile fails due to being unauthorized. It may also be that we aren't requeueing when there is a validation failure. Status conditions: status: conditions: - category: Critical lastTransitionTime: "2021-03-31T14:14:32Z" message: 'Test connect failed. ' reason: ConnectFailed status: "True" type: TestConnectFailed - category: Critical lastTransitionTime: "2021-03-30T17:00:05Z" message: 'Reconcile failed: [Unauthorized]. See controller logs for details.' status: "True" type: ReconcileFailed observedDigest: 3bdd8189de081b4315990a49028a49b7df9b6dd503490774848d619eea48dfe8 operatorVersion: latest registryPath: docker-registry-default.apps.djwocp3a.mg.dog8code.com
Seems to be related to https://bugzilla.redhat.com/show_bug.cgi?id=1741281, but in this case I didn't see recovery happen.
This is something I'm pretty confident I often did in the past without issue. I'm wondering if something potentially regressed here?
Jaydip wasn't able to reproduce this issue. We're waiting on additional reproducer steps from Xin in a related BZ, but this may already be fixed, could be a UI regression that has since been resolved. Will reach out to Ian to see if he knows anything.
I think this bug is fixed, I can no longer reproduce the issue. Guessing Pranav resolved it when doing testing on change of SA token for client caching work in https://github.com/konveyor/mig-controller/pull/1037, and Ian may have fixed UI related parts around waiting to update status of MigCluster until changes are made in modal.
Verified using MTC 1.5.0 openshift-migration-rhel7-operator@sha256:00e77706ca22bcb557d13c16822180fc877e6ea1639a72fda8eb9f5488b039a2 - name: MIG_CONTROLLER_REPO value: openshift-migration-controller-rhel8@sha256 - name: MIG_CONTROLLER_TAG value: 7f657df15e9514df4ef42da3431f558a19b8d3233a2ef1222cd8e27793c93816 After breaking the migcluster connection I was able to fix it by editing the SA token and the migcluster resource returned to Ready status. Moved to VERIFIED.
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 (Migration Toolkit for Containers (MTC) image release advisory 1.5.0), 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/RHEA-2021:2929