gcs Skeleton E2E

As a first step at getting an e2e suite for GCS, we should have an automated job that (nightly?):

Brings up a kube cluster using this repo
Deploys GCS on that cluster
Verifies that the expected pods started
- gluster (x3)
- csi for gluster file (attacher, provisioner, node x3)

Oct 15 '18 18:10 JohnStrunk

In the weekly status meeting, it was decided this should also start a Job to verify we can read/write to Gluster.

Oct 18 '18 19:10 JohnStrunk

@nigelbabu @JohnStrunk I believe we're not yet ready to get this rolling for GCS 0.2? If so can one of you move it to GCS/0.3?

Oct 25 '18 12:10 atinmu

@nigelbabu ping! where're we on this?

Nov 13 '18 11:11 atinmu

I'm working on deploying GCS on CentOS CI. Currently looking through the deploy scripts to fill in any gaps, including host setup on CentOS CI. I expect the outcome of this issue to be a reproducable deployment that could be torn down completely and brought up fully to a known state against which tests from anteater could be run. Expect an update on Monday, 26th.

Nov 22 '18 03:11 brainfunked

The CentOS CI job is green. However, there is a possibility that it doesn't always work reliably due to factors such as a delay in the network response and thus a timeout in the test. I'm now investigating the reliability of the deployments.

Dec 03 '18 08:12 brainfunked

@brainfunked @nigelbabu Should this issue stay as part of GCS/0.5 label?

Jan 07 '19 05:01 atinmu

Push it to 0.6, because this will be correctly fixed in gluster/gluster-csi-driver#136

Jan 09 '19 03:01 nigelbabu