Description of problem: https://github.com/fusor/mig-controller/issues/409 Depending on the access mode of the volumes, it's possible for the restore to fail if a stage pod is still terminating at restore time because the volume will have a lock on it. To fix this, the controller should be waiting for all stage pods to finish terminating before running the restore. Version-Release number of selected component (if applicable): 4.2.0 How reproducible: 70%? Hard to say since it's a race condition, but with specific volumes it's more reproducible than others. Steps to Reproduce: 1. Run a migration with RWO volumes Actual results: Final restore fails due to being unable to mount the volume Expected results: Restore succeeds and restore doesn't run until stage pods are finished terminating. Additional info:
https://github.com/konveyor/mig-controller/pull/444
Verified using CAM 1.2 stage Once the stage pods are created, there is a phase to wait for the stage pods to finish before restoring. And the execution is blocked there until stage pods are removed (after "Terminating" status) message: 'Step: 19/31' reason: StageRestoreCreated message: 'Step: 21/31' reason: EnsureStagePodsTerminated status: "True" message: 'Step: 27/31' reason: FinalRestoreCreated status: "True"
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, 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-2020:2326