resouces: Add readinessProbe for ctdb container

[spuiuk]

Use samba-container check ctdb-nodestatus as readiness probe for the ctdb container.

Signed-off-by: Sachin Prabhu sprabhu@redhat.com

[spuiuk]

To test, I disabled ctdb on one of the running pods.

$ kubectl exec -it smbshare3-0 -c ctdb -- /bin/bash
[root@smbshare3-0 /]# ctdb nodestatus
pnn:0 10.244.1.14      OK (THIS NODE)
[root@smbshare3-0 /]# ctdb disable
[root@smbshare3-0 /]# ctdb nodestatus
pnn:0 10.244.1.14      DISABLED (THIS NODE)

And can confirm that the readiness probe fails by looking at

$ kubectl get pods -w 
NAME                               READY   STATUS    RESTARTS   AGE
samba-ad-server-86b7dd9856-8wj9q   1/1     Running   0          10m
smbshare3-0                        3/3     Running   0          2m54s
smbshare3-1                        3/3     Running   0          111s
smbshare3-0                        2/3     Running   0          3m52

and

$ kubectl describe pod smbshare3-0
...
  Warning  Unhealthy               2m50s (x12 over 4m30s)  kubelet                  Readiness probe failed: pnn:0 10.244.1.14      DISABLED (THIS NODE)

Checking the spec of the statefulset created.

$ kubectl edit statefulset smbshare3 
...
        - name: SAMBACC_CTDB
          value: ctdb-is-experimental
        image: quay.io/samba.org/samba-server:latest
        imagePullPolicy: Always
        name: ctdb
        readinessProbe:
          exec:
            command:
            - samba-container
            - check
            - ctdb-nodestatus
          failureThreshold: 3
          periodSeconds: 10
          successThreshold: 1
          timeoutSeconds: 1
...

[anoopcs9]

/test all