Scott Seago

But again, this may all just work once you get the pod duplication fix. One thing to try -- if you restore the pod but not the Deployment, you won't...

@bluzarraga ok, so I guess that's good and bad news. The good news is that the ReplicaSet/pod duplication issue is not causing the network problem and the pod issue will...

@reasonerjt The second one is the longer-term sticking point. If we removed enableCSI we'd need to replace it with something like useCSI which would do the same thing.

I see that this has all been refactored on main since 1.12: https://github.com/vmware-tanzu/velero/pull/7130 So we still have logic to cancel Accepted DUs and DDs are still canceled on start: ```...

To fully fix this, we may need to handle hooks after async actions have completed. One possibility: attempt hook initially with small timeout for Pod readiness, and if pod is...

@blackpiglet "but I don't get why that will make the volumes restore not work in parallel." -- if the timeout for the hook is long enough to complete the download,...

@allenxu404 Actually, the BSL is the backupstoragelocations CR -- backuprepositories is used for kopia/restic for FS backup.

It sounds like we need to completely step back. @shubham-pampattiwar has been designing for one set of requirements, but the other reviews seem to have a completely different set of...

@weshayutin "3. Finally if there is no match for a volumePolicy and no match for opt-in/out method, the velero server install default (FSBackupOrSnapShot) will be authoritative." This is not quite...

@asaf-erlich If your goal is to back up with a mixture of fs-backup and datamover, that is absolutely supported by this design. Policies will identify volumes for fs-backup and snapshot,...