feat: implement snapshots as tarballs

[wozniakjan]

What type of PR is this?

/kind feature

What this PR does / why we need it:
This PR proposes simple support for VolumeSnapshots as tarballs created with

tar -C /[src_vol_path] -czvf /[snapshot_path]/[snapshot_name]/[src_vol_name].tar.gz .

The reason why [src_vol_uuid] is included in the snapshot path is to be able to efficiently discover AlreadyExists cases

It is the second out of three intended PRs to provide basic snapshot workflow including volume cloning from snapshots. It is preceded by #426 which implemented volume cloning from other volumes.

Which issue(s) this PR fixes:

Moving forward with snapshots as discussed in #31

Special notes for your reviewer:
Some parts of this PR related to helm chart and external-snapshotter are not ideal. As a followup task, I would like to attempt to address kubernetes-csi/external-snapshotter#812. That way all csi-driver helm charts including this one, instead of embedding external-snapshotter CRDs and controller resources and duplicating them, they could just add the official external-snapshotter chart as a dependency. This should also simplify updating to new releases. But not in scope to be addressed here.

Does this PR introduce a user-facing change?:

feat: volume snapshots

[coveralls]

Pull Request Test Coverage Report for Build 4597662542

• 173 of 218 (79.36%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+1.1%) to 78.426%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pkg/nfs/controllerserver.go	172	217	79.26%

Totals
Change from base Build 4594162046:	1.1%
Covered Lines:	847
Relevant Lines:	1080

---

💛 - Coveralls

[andyzhangx]

when the main work of this feature is almost done, pls also add some unit tests, ut coverage is decreasing a lot

Overall coverage decreased (-9.8%) to 63.45%

[wozniakjan]

I will turn this into draft for now and undraft when it's reviewable again

I will make this as WIP for now and un-WIP when it's reviewable again

it's reviewable again :)

[wozniakjan]

hey @andyzhangx, any ideas why external-e2e tests continue to fail with

log

        Mar 20 16:37:35.445: Unexpected error:
            <*errors.StatusError | 0xc003ea4320>: {
                ErrStatus: {
                    TypeMeta: {Kind: "", APIVersion: ""},
                    ListMeta: {
                        SelfLink: "",
                        ResourceVersion: "",
                        Continue: "",
                        RemainingItemCount: nil,
                    },
                    Status: "Failure",
                    Message: "the server could not find the requested resource",
                    Reason: "NotFound",
                    Details: {
                        Name: "",
                        Group: "",
                        Kind: "",
                        UID: "",
                        Causes: [
                            {
                                Type: "UnexpectedServerResponse",
                                Message: "404 page not found",
                                Field: "",
                            },
                        ],
                        RetryAfterSeconds: 0,
                    },
                    Code: 404,
                },
            }
            the server could not find the requested resource
        occurred

        test/e2e/storage/framework/snapshot_resource.go:65

• https://prow.k8s.io/view/gs/kubernetes-jenkins/pr-logs/pull/kubernetes-csi_csi-driver-nfs/430/pull-csi-driver-nfs-external-e2e/1637849095554994176
• https://storage.googleapis.com/kubernetes-jenkins/pr-logs/pull/kubernetes-csi_csi-driver-nfs/430/pull-csi-driver-nfs-external-e2e/1637849095554994176/build-log.txt

I think the failure is here and iiuc my guess could be that SnapshotClass maybe doesn't exist in the kube-apiserver or maybe the create call sends an empty GVK in the request?
https://github.com/kubernetes/kubernetes/blob/v1.24.0/test/e2e/storage/framework/snapshot_resource.go#L64-L65

maybe this bit is not quite right?

csi-driver-nfs/test/external-e2e/testdriver.yaml

Lines 6 to 7 in 826d2f3

	SnapshotClass:
	FromName: true

EDIT: for future reference, fixed in 5643e5f and fd1f166, the external-e2e-tests install a helm chart of this driver and expect external-snapshotter CRDs and controller to be part of the chart.

[wozniakjan]

after other PRs merged, there was a go.mod conflict, appended go mod tidy to the PR after rebase

[andyzhangx]

after other PRs merged, there was a go.mod conflict, appended go mod tidy to the PR after rebase

@wozniakjan could you squash the commits to only a few commits?

[wozniakjan]

@wozniakjan could you squash the commits to only a few commits?

Thanks for pinging, rebased and did the best effort to logically squash commits, ptal @andyzhangx

[smuda]

Thank you for adding this feature! Will there be a status or event I can listen for to be sure the snapshot has succeeded?

The scenario is shutting down a database pod (to flush to disk), requesting a snapshot and then waiting for it to finish before starting the database pod again.

[wozniakjan]

Will there be a status or event I can listen for to be sure the snapshot has succeeded?

the VolumeSnapshot has status.readyToUse field which is set to true by external-snapshotter controller after the snapshot routine in a specific driver has completed
https://github.com/kubernetes-csi/external-snapshotter/blob/129de4cbb3fc5679285d70d191e66976481082a9/client/config/crd/snapshot.storage.k8s.io_volumesnapshots.yaml#L167

And just like you are used to with other kubernetes resources, you can run

$ kubectl describe volumesnapshot [name]
$ kubectl describe volumesnapshotcontent [name]

to debug for any errors

[andyzhangx]

/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: andyzhangx, wozniakjan

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [andyzhangx]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment