Description of problem: Currently the stage pods are hardcoded to pull from registry.access.redhat.com, which has been deprecated in favor of registry.redhat.io. This has been raised as a problem by a customer who only has whitelisted registry.redhat.io; they are unable to pull from registry.access.redhat.com. https://www.redhat.com/en/blog/transitioning-red-hat-container-registry Relevant controller source: https://github.com/konveyor/mig-controller/blob/master/pkg/controller/migmigration/stage.go#L571 https://github.com/konveyor/mig-controller/blob/master/pkg/controller/migmigration/stage.go#L610 Version-Release number of selected component (if applicable): 1.2.4 How reproducible: Every time
Looks like registry.redhat.io/rhel7 also exists (if we can maybe we should switch to ubi8, but that's a separate discussion). Perhaps we pass the REGISTRY env var we set on the operator to the controller container and use that as the registry prefix here? CC'ing Scott and Derek for their input since this is more their expertise.
PRs in progress: https://github.com/konveyor/mig-controller/pull/628 https://github.com/konveyor/openshift-migration-plugin/pull/69 https://github.com/konveyor/mig-operator/pull/400
https://github.com/konveyor/mig-controller/pull/628 https://github.com/konveyor/openshift-migration-plugin/pull/69 https://github.com/konveyor/mig-operator/pull/400
Verified using CAM 1.2.5 stage openshift-migration-rhel7-operator@sha256:493e56a8b609eadad79d2df8e3c6402b2014ebcb645a939fa34004b056ea1a2e We have configured in the migrationcontroller resource this values for stage and registry pods migration_registry_image: docker.io/registry migration_registry_repo: registry migration_registry_version: 2 migration_stage_version: 44f0362c8570d707582bd428aaf18f390ce915ef72cdeb60cf2699171dbda3c8 (non default one) velero_restic_restore_helper_version: 44f0362c8570d707582bd428aaf18f390ce915ef72cdeb60cf2699171dbda3c8 (non default one) We did it in target cluster, and in source cluster, and we could verify that the values used were the right ones in the right clusters, that the migration-cluster-config configmap was updated and that everything was working fine in migrations with pvcs and internal images (cakephp application). We move the status to VERIFIED.
We verified too that the file /operator.yml inside the operator pod points to the right registry registry.redhat.io and that this variables define the content of the configmap and the image used for the stage pod migration_stage_image: XXXXXXX migration_stage_repo: XXXXXX migration_stage_version: XXXXXXXXX
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 (Cluster Application Migration (CAM) Tool Image Release Advisory 1.2.5), 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-2020:3535