fix validate-cluster.sh

[caesarxuchao]

attempt to fix #47379.

Without this fix, the validate-cluster.sh never retries if kubectl-retry get cs fails.

cc @dchen1107

[gmarek]

This is wrong. return_value is used in exit call, which expects integer, plus you shouldn't overwrite what's currently stored in return_value.

IIUC your main goal is to to prevent errexit to kill the script when get cs fails. If that's the case you should write:

echo $(kubectl_retry get cs || true)

[caesarxuchao]

Yes. Thanks.

[caesarxuchao]

/release-note-none

[krousey]

/lgtm

[caesarxuchao]

@gmarek @krzyzacy could you take another look? Thanks.

[krzyzacy]

/lgtm

[krzyzacy]

/assign @mikedanese

[mikedanese]

/approve

[marun]

/retest

[dchen1107]

Can we run validation_cluster.sh with the new change against the existing k8s cluster to verify the change? upgrade.sh is flaky. Without any change, the new run of
both master and nodes are successfully upgraded to 1.7 in a new run: https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/ci-kubernetes-e2e-gce-1-6-1-7-upgrade-cluster/55?log#log

I0620 10:54:17.484] == Finished upgrading nodes to v1.7.0-beta.2.120+c8cd7f318aea33. ==
I0620 10:54:17.484] == Validating cluster post-upgrade ==
I0620 10:54:17.519] Validating gce cluster, MULTIZONE=
I0620 10:54:17.520] Project: gce-up-c1-3-g1-4-up-clu
I0620 10:54:17.520] Zone: us-central1-f
I0620 10:54:18.571] Found 4 node(s).
I0620 10:54:18.774] NAME                              STATUS                     AGE       VERSION
I0620 10:54:18.774] bootstrap-e2e-master              Ready,SchedulingDisabled   13m       v1.7.0-beta.2.120+c8cd7f318aea33
I0620 10:54:18.774] bootstrap-e2e-minion-group-4jm4   Ready                      8m        v1.7.0-beta.2.120+c8cd7f318aea33
I0620 10:54:18.774] bootstrap-e2e-minion-group-gw1d   Ready                      4m        v1.7.0-beta.2.120+c8cd7f318aea33
I0620 10:54:18.774] bootstrap-e2e-minion-group-km98   Ready                      23s       v1.7.0-beta.2.120+c8cd7f318aea33
I0620 10:54:19.486] Validate output:
I0620 10:54:20.191] NAME                 STATUS    MESSAGE              ERROR
I0620 10:54:20.191] controller-manager   Healthy   ok                   
I0620 10:54:20.191] scheduler            Healthy   ok                   
I0620 10:54:20.191] etcd-0               Healthy   {"health": "true"}   
I0620 10:54:20.191] etcd-1               Healthy   {"health": "true"}   
I0620 10:54:20.195] Cluster validation succeeded
I0620 10:54:20.195] == Post-Upgrade Node OS and Kubelet Versions ==
I0620 10:54:20.395] name: "bootstrap-e2e-master", osImage: "Container-Optimized OS from Google", kubeletVersion: "v1.7.0-beta.2.120+c8cd7f318aea33"
I0620 10:54:20.395] name: "bootstrap-e2e-minion-group-4jm4", osImage: "Debian GNU/Linux 7 (wheezy)", kubeletVersion: "v1.7.0-beta.2.120+c8cd7f318aea33"
I0620 10:54:20.395] name: "bootstrap-e2e-minion-group-gw1d", osImage: "Debian GNU/Linux 7 (wheezy)", kubeletVersion: "v1.7.0-beta.2.120+c8cd7f318aea33"
I0620 10:54:20.395] name: "bootstrap-e2e-minion-group-km98", osImage: "Debian GNU/Linux 7 (wheezy)", kubeletVersion: "v1.7.0-beta.2.120+c8cd7f318aea33"

[dchen1107]

/lgtm

[k8s-github-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: caesarxuchao, dchen1107, krousey, krzyzacy, mikedanese

Associated issue: 47379

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

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

• cluster/OWNERS [dchen1107,mikedanese]

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

[dchen1107]

@caesarxuchao and I did validation this change against a non-exist cluster. It works as expected. Thanks!

[caesarxuchao]

/test pull-kubernetes-federation-e2e-gce

[k8s-github-robot]

Automatic merge from submit-queue