Whitelist kubemark in node_ssh_supported_providers for log dump

[shyamjvs]

Seems like PR #41739 did not truly enable hollow node logging, as there were a couple of issues:

• kubemark is not whitelisted as a k8s provider for which we can perform node SSH (because that's the case with hollow nodes), but we can still SSH to real nodes. So added it to node_ssh_supported_providers
• The if [[ "${KUBERNETES_PROVIDER}" -eq "kubemark" && "${ENABLE_HOLLOW_NODE_LOGS:-}" -eq "true" ]]; in bash didn't work as I expected. Fixed it now and checked.

With these changes, it should work now.

cc @kubernetes/sig-scalability-misc @wojtek-t @gmarek

[k8s-reviewable]

This change is 

[wojtek-t]

/lgtm
/approve

[k8s-github-robot]

[APPROVALNOTIFIER] This PR is APPROVED

The following people have approved this PR: shyamjvs, wojtek-t

Needs approval from an approver in each of these OWNERS Files:

• cluster/OWNERS [wojtek-t]

You can indicate your approval by writing /approve in a comment
You can cancel your approval by writing /approve cancel in a comment

[yujuhong]

By the way, the original PR broke the log collection for all builds. We no longer collect any logs.

I0221 04:18:09.992] Dumping master and node logs to /workspace/_artifacts
I0221 04:18:09.992] /workspace/kubernetes/cluster/log-dump.sh: line 132: gce: unbound variable
I0221 04:18:09.993] /workspace/kubernetes/cluster/log-dump.sh: line 132: gce: unbound variable
I0221 04:18:09.993] /workspace/kubernetes/cluster/log-dump.sh: line 132: gce: unbound variable
I0221 04:18:09.993] /workspace/kubernetes/cluster/log-dump.sh: line 132: gce: unbound variable
I0221 04:18:09.993] /workspace/kubernetes/cluster/log-dump.sh: line 132: gce: unbound variable

Bumping the priority of this fix.

[yujuhong]

/cc @k8s-oncall

[shyamjvs]

@yujuhong Thanks for bringing this up. This PR should also fix that issue. It was a problem in this line where I used '-eq' comparator instead of '==' for strings.

[shyamjvs]

Merging this PR manually to fix broken log dumping for tests.

[yujuhong]

@shyamjvs, I think it's good to loop the oncall in, and let them do the merge if possible :-)

[shyamjvs]

Sure. I'll take care of this in future. :)
In this case, I thought that fixing the logs is crucial to debugging test failures that might happen in the meanwhile.