external-snapshotter deployment issues

[saad-ali]

Only the dev deployment script deploys an external-snapshotter.

gcp-compute-persistent-disk-csi-driver/deploy/kubernetes/overlays/dev/controller_images.yaml

Lines 13 to 15 in 191be85

	- name: csi-snapshotter
	imagePullPolicy: Always
	image: quay.io/k8scsi/csi-snapshotter:v0.4.0

1. Shouldn't the stable deployment script also deploy the external snapshotter (at least as a flag controlled option)?

2. The external-snapshotter in dev is out of date and results in the controller crash looping:

  csi-snapshotter:
    Container ID:  docker://2c24cd5a64a38d1bfebf898cf8779070184e541c90c3ce7204d313aba893c863
    Image:         quay.io/k8scsi/csi-snapshotter:v0.4.0
    Image ID:      docker-pullable://quay.io/k8scsi/csi-snapshotter@sha256:e2cc9f3e61d53e4fd35b96c103bcea48ca41ef3b38662f94e9da749f359ba130
    Port:          <none>
    Host Port:     <none>
    Args:
      --v=5
      --csi-address=$(ADDRESS)
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Error
      Exit Code:    1
      Started:      Wed, 19 Dec 2018 22:03:11 -0800
      Finished:     Wed, 19 Dec 2018 22:03:11 -0800
    Ready:          False
    Restart Count:  5

It is crash looping because the external-snapshotter version is wrong:

$ kubectl logs csi-gce-pd-controller-0 -c csi-snapshotter
I1220 06:05:52.112139       1 main.go:76] Version: v0.4.0-4-ge6d7b9ac
I1220 06:05:52.144724       1 connection.go:83] Connecting to /csi/csi.sock
I1220 06:05:52.144908       1 connection.go:111] Still trying, connection is CONNECTING
I1220 06:05:52.145223       1 connection.go:108] Connected
I1220 06:05:52.145248       1 connection.go:259] GRPC call: /csi.v0.Identity/GetPluginInfo
I1220 06:05:52.145672       1 connection.go:261] GRPC response: 
I1220 06:05:52.145721       1 connection.go:262] GRPC error: rpc error: code = Unimplemented desc = unknown service csi.v0.Identity
E1220 06:05:52.145746       1 main.go:131] rpc error: code = Unimplemented desc = unknown service csi.v0.Identity

The external-snapshotter version should be updated to point to a CSI 1.x version.

3. external-snapshotter only exists on quay not on gcr.io/gke-release. We should have external-snapshotter deployed on gcr.io/gke-release.

CC @jingxu97 @davidz627

[msau42]

1. Let's create a new "alpha" config to deploy the snapshotter container.
2. This should be updated to use snapshotter v1.0.x
3. We need to create new internal flows to build snapshotter. Will sync up with jing offline

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[davidz627]

resolved with multiple PRs.

Snapshotter has been updated, and there is now an alpha config with the external snapshotter
/close

[k8s-ci-robot]

@davidz627: Closing this issue.

In response to this:

resolved with multiple PRs.

Snapshotter has been updated, and there is now an alpha config with the external snapshotter
/close

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/test-infra repository.