Created attachment 1626028 [details] log files, bucket backup and bucket restore-logs Description of problem: When a migration from 3.7 to 4.2 is performed, cronjobs are not migrated. Version-Release number of selected component (if applicable): OCP4 NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.2.0-0.nightly-2019-10-11-230724 True False 33h Cluster version is 4.2.0-0.nightly-2019-10-11-230724 OCP3 oc v3.7.126 kubernetes v1.7.6+a08f5eeb62 features: Basic-Auth GSSAPI Kerberos SPNEGO Server https:// openshift v3.7.126 kubernetes v1.7.6+a08f5eeb62 Controller Image: openshift-migration-controller-rhel8:v1.0 openshift-migration-controller-rhel8@sha256:46c622e0fbe64165b09930738a7d111c875976b54c8236ddce328cb5470d60ab Velero Images: openshift-migration-velero-rhel8:v1.0 openshift-migration-velero-rhel8@sha256:bc0de9d8edcec4d2ca111c8f73e52a558e44619e2a592aefff7224895a55b4d0 openshift-migration-plugin-rhel8:v1.0 openshift-migration-plugin-rhel8@sha256:3f0f4c2efb27788f74b72a4bc1f526dc5b4637a67478030df64e5d792fe99511 How reproducible: Always Steps to Reproduce: 1.Create a namespace oc new-project cronjob-test 2.Create an internal image oc import-image intalpine:int --from=docker.io/alpine:latest --confirm 3.Create a cronjob using a external image oc process -f https://raw.githubusercontent.com/sergiordlr/temp-testfiles/master/app_migration/cronjob/3.7/hello_cron_template.yml -p CRON_NAME=external-img -p IMAGE=docker.io/alpine | oc create -f - 4.Create a cron that uses this internal image oc process -f https://raw.githubusercontent.com/sergiordlr/temp-testfiles/master/app_migration/cronjob/3.7/hello_cron_template.yml -p CRON_NAME=internal-img -p IMAGE=docker-registry.default.svc:5000/cronjob-test/intalpine:int | oc create -f - cronjob.batch/internal-img created 5. Migrate the namespace Actual results: The external image imagestream is migrated. But the cronjobs are not. $ oc get cronjobs -n cronjob-test No resources found. Expected results: Both cronjobs should be migrated and work fine in the OCP4 target cluster. Additional info: All logs are included in the attached file, as well as the bucket backup and the bucket restore-logs files.
Our understanding of the problem is that the source is using a CronJob version of "batch/v2alpha1" This has changed and is no longer available in the OCP 4 destination cluster. We don't have a means of identifying or fixing this issue yet, planning to look into how we can detect these problems and fix in future. For now moving this bug to a future release as it won't make our CAM 1.1.0 release (timeframe of OCP 4.3.).
Tracking this as a larger feature for allow auto conversion of older api versions to newer/supported versions. See: https://issues.redhat.com/browse/MIG-168
Closing as this should be captured by the larger RFE for allowing auto conversion of auto apis.