[Feature request] Add support for custom path for snapshots
smuda opened this issue · 6 comments
Is your feature request related to a problem?/Why is this needed
When creating snapshots they are created at the NFS root, with names that is unclear what volume claim that is the origin of the snapshot. It's also problematic after hundreds of snapshots to have them all in the same folder as the volumes.
Describe the solution you'd like in detail
For the volumes I can specify the subpath in StorageClass.parameters.subdir with templating, for example ${pvc.metadata.namespace}/pvc-${pvc.metadata.name}
I'd propose using the same model for VolumeSnapshotClass.parameters.subdir with support for at least vs.metadata.name, vs.metadata.namespace and vs.spec.source.persistentVolumeClaimName.
@smuda the *.tar.gz already contains the pv name, check below logs, is that enough?
[pod/csi-test-controller-9784fb589-nprbv/nfs] I1213 02:44:38.584541 1 utils.go:108] GRPC call: /csi.v1.Controller/CreateSnapshot
[pod/csi-test-controller-9784fb589-nprbv/nfs] I1213 02:44:38.584560 1 utils.go:109] GRPC request: {"name":"snapshot-a56cdc8b-a0e2-4f46-b3a0-916ffed068eb","source_volume_id":"nfs-server.default.svc.cluster.local##pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac##"}
[pod/csi-test-controller-9784fb589-nprbv/nfs] I1213 02:44:38.618522 1 nodeserver.go:144] volume(nfs-server.default.svc.cluster.local##pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac##) mount nfs-server.default.svc.cluster.local:/ on /tmp/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac succeeded
[pod/csi-test-controller-9784fb589-nprbv/nfs] I1213 02:44:38.618649 1 controllerserver.go:357] archiving /tmp/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac -> /tmp/snapshot-a56cdc8b-a0e2-4f46-b3a0-916ffed068eb/snapshot-a56cdc8b-a0e2-4f46-b3a0-916ffed068eb/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac.tar.gz
[pod/csi-test-controller-9784fb589-nprbv/nfs] I1213 02:44:38.622663 1 controllerserver.go:362] archived /tmp/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac -> /tmp/snapshot-a56cdc8b-a0e2-4f46-b3a0-916ffed068eb/snapshot-a56cdc8b-a0e2-4f46-b3a0-916ffed068eb/pvc-9e6bdd5a-c6c0-466a-b0b5-9b6cbf84a0ac.tar.gz
Only as long as the cluster is healthy :-)
So the PVC we create have a nice, understandable name which, especially in the context of a namespace, makes sense:
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS
alert-pg Bound pvc-cc3f3d08-f0ef-4769-a9c7-bf0901b108eb 5Gi RWO nfs-csi
...
But the automatically created PVs does not get understandable names:
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON
pvc-cc3f3d08-f0ef-4769-a9c7-bf0901b108eb 5Gi RWO Delete Bound bd/alert-pg nfs-csi
...
So even though the PV name ("pvc-cc3f3d08-f0ef-4769-a9c7-bf0901b108eb") is included in the path, it doesn't really tell me anything (not even namespace) when trying to restore the PVs after the cluster has gone "boom".
In my context, I'd like to create the snapshots inside a directory for the namespace with a name taken from the PVC, and of course some kind of uid (or date/time?) to distinguish between the snapshots, for example:
subdir="snapshots/${vs.metadata.namespace}/${vs.spec.source.persistentVolumeClaimName}"
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle rotten - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Reopen this issue with
/reopen - Mark this issue as fresh with
/remove-lifecycle rotten - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/close not-planned
@k8s-triage-robot: Closing this issue, marking it as "Not Planned".
In response to this:
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied- After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied- After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closedYou can:
- Reopen this issue with
/reopen- Mark this issue as fresh with
/remove-lifecycle rotten- Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/close not-planned
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.