This bug was initially created as a copy of Bug #1897891 I am copying this bug because: the CLI user should be informed with an error about that the import fails due to missing vddk-init image Description of problem: There is no information regarding vddk-init image not found in ‘v2v-vmware’ configmap An error can be found in cdi-deployment pod log: "level":"error","ts":1605106366.84607,"logger":"controller","msg":"Reconciler error","controller":"import-controller","name":"v2v-rhel8-foreman-os-harddisk1","namespace":"default","error":"Found v2v-vmware ConfigMap in namespace openshift-cnv, but it does not contain a 'vddk-init-image' entry." Version-Release number of selected component (if applicable): OCP 4.6.3 CNV 2.5.0 How reproducible: 100% Expected results: An import error message to let the user know that the migration is blocked due the missing of vddk-init image
Sam, can you bubble up the error in the VirtualMachineImport CR?
Unfortunately not without changes to CDI first. The reconcile error indicating that the vddk init image is missing only appears in the CDI controller log. CDI treats this as a transient reconcile error, because as soon as the the vddk-init-image key is added to the configmap the importer pod can be created. It's not propagated up to the DataVolume status or anywhere else. From the perspective of the DataVolume (and VMIO), it's just waiting a long time for the importer pod to be created.
Documentation says it should be configured. Moving to '2.7.0'.
We'll handle that in MTV when VMIO logic is absorbed. Closing.