Michael Steven Fruchtman

It is the responsibility of the external-snapshotter. https://github.com/kubernetes/enhancements/tree/master/keps/sig-storage/3141-prevent-volume-mode-conversion `With this change, the controller will fetch the Spec.PersistentVolumeMode of the PV and add that to newly introduced Spec.SourceVolumeMode field of the...

Ceph-csi has already updated to external-snapshotter v8.0.0 in development https://github.com/ceph/ceph-csi/commit/c48f5bff371362796421b2f8c2b265476c9e9ed7 There are no existing releases with external-snapshotter v8.0.0.

I was finally given access to the environment and found the issue was an _extraordinarily_ downlevel install of Rook.io v1.8.8. Internal procedures were not followed. The installer simply continued to...

Closing issue since reversion has been merged.

@blackpiglet Another unusual issue along similar lines was proxy support. I don't know the software's name but the user was requesting how to add Pod labels to any Pod that...

That would work. I'm very flexible on whatever solution anyone proposes. We just need a way to make sure to set DataUpload and DataDownload associated Pods labels and annotation. Unlike...

Yes, I noticed. Which is why I didn't say it blocked scheduling, it only skewed it. The existing state model allows datamovers to remain in state Pending without issue when...

@Lyndon-Li I've thought of another thing that might help with scheduling besides PriorityClass and priority options. Currently the node-agent-config podResources have to be set sufficiently high to account for the...

> > Currently the node-agent-config podResources have to be set sufficiently high > > Actually, the default behavior is `BestEfforts`, that is no request/limit is set. And this is also...

I have another reason to prefer the first solution of only creating datamover Pods equal to the loadConcurrency. Though this is admittedly NOT a Velero problem. Some recent investigation (this...