vmware-tanzu / velero

Backup and migrate Kubernetes applications and their persistent volumes
https://velero.io
Apache License 2.0
8.79k stars 1.41k forks source link

Cluster Migration with Data Movement #8270

Open raynay-r opened 1 month ago

raynay-r commented 1 month ago

Discussed in https://github.com/vmware-tanzu/velero/discussions/8194

Originally posted by **raynay-r** September 8, 2024 Hi! I have two OKD cluster and want to migrate an application between them. Both clusters run rook-ceph with the same storage classes. I am using an additional Minio instance as backup storage. Both OKD cluster have access to the Minio instance. I was able to successfully migrate the application using the file system backup, but I am not able to get it working with data movement. The backup itself works without issue but when I try to restore in the second cluster I get the following error ``` Errors: Velero: Cluster: Namespaces: projects-xwiki-restore: error preparing persistentvolumeclaims/projects-xwiki-restore/data-xwiki-mysql-0: rpc error: code = Unknown desc = fail get DataUploadResult for restore: xwiki-restore-23: no DataUpload result cm found with labels velero.io/pvc-namespa ce-name=projects-xwiki.data-xwiki-mysql-0,velero.io/restore-uid=f7c9dc29-e46e-4b9f-b67c-46564c16c0d9,velero.io/resource-usage=DataUpload ``` Should Data Movement between clusters work?

Opening an issue from a discussion, because I got no further answers in the discussions section and I was told that this should work. Please take a look at the linked discussion for detailed description and log output.

Would be great if you could help me out here.

Thanks!

Lyndon-Li commented 1 month ago

Could you share velero log bundle by running velero debug?

raynay-r commented 1 month ago

Sure. Here are the bundles from the source and target cluster:

bundle-source-cluster.tar.gz bundle-target-cluster.tar.gz

ywk253100 commented 1 month ago

The restore forgejo-restore-1 in the debug bundle was created in 2024-09-14, but the log started from 2024-10-03, cannot find any useful information related to the failed restore. Could you create a new restore and collect the debug bundle again?

raynay-r commented 1 month ago

I created a new backup with the name "forgejo-backup-10" on the source cluster and a restore with the name "forgejo-restore-11" on the target cluster an could reproduce the same error again. Here are the bundles: bundle-target-cluster.tar.gz bundle-source-cluster.tar.gz

raynay-r commented 4 weeks ago

@ywk253100 @Lyndon-Li

Did you find anything in the debug bundles?