Ability to resize PVCs used in a PravegaCluster

[vedanthh]

Increasing Journal,Cache and ledgers size not reflecting for new/scaledup bookie and segmentstore using kubectl edit

Snip:

kubectl edit PravegaCluster nautilus -n nautilus-pravega
    storage:
      journalVolumeClaimTemplate:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 500Gi -- increased from 250 Gi
      ledgerVolumeClaimTemplate:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 500Gi -- increased from 250Gi
  pravega:
    cacheVolumeClaimTemplate:
      accessModes:
      - ReadWriteOnce
      resources:
        requests:
          storage: 200Gi -- increased from 100 Gi

[root@manager1 ~]# kubectl get pvc --all-namespaces
NAMESPACE          NAME                                            STATUS    VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE

nautilus-pravega   cache-nautilus-segmentstore-3                   Bound     pvc-49da332c-f2de-11e8-b82a-005056bd38d8   100Gi      RWO            standard       58m
nautilus-pravega   cache-nautilus-segmentstore-4                   Bound     pvc-b67443d3-f2e0-11e8-b82a-005056bd38d8   100Gi      RWO            standard       41m
nautilus-pravega   cache-nautilus-segmentstore-5                   Bound     pvc-b67443d3-f2e0-11e8-b82a-005056bd38d8   100Gi      RWO            standard       42m

nautilus-pravega   journal-nautilus-bookie-3                       Bound     pvc-49d3457c-f2de-11e8-b82a-005056bd38d8   250Gi      RWO            standard       58m
nautilus-pravega   journal-nautilus-bookie-4                       Bound     pvc-f1ee7509-f2df-11e8-b82a-005056bd38d8   250Gi      RWO            standard       46m
nautilus-pravega   journal-nautilus-bookie-5                       Bound     pvc-32263446-f2e0-11e8-b82a-005056bd38d8   250Gi      RWO            standard       45m

nautilus-pravega   ledger-nautilus-bookie-3                        Bound     pvc-49d3fff1-f2de-11e8-b82a-005056bd38d8   250Gi      RWO            standard       58m
nautilus-pravega   ledger-nautilus-bookie-4                        Bound     pvc-f1ef09f7-f2df-11e8-b82a-005056bd38d8   250Gi      RWO            standard       46m
nautilus-pravega   ledger-nautilus-bookie-5                        Bound     pvc-32268e3a-f2e0-11e8-b82a-005056bd38d8   250Gi      RWO            standard       45m

[root@manager1 ~]# kubectl -n nautilus-pravega exec -it nautilus-bookie-5 -- /bin/sh
/opt/bookkeeper # df -h
Filesystem                Size      Used Available Use% Mounted on
overlay                   9.7G      5.3G      3.9G  58% /
tmpfs                    64.0M         0     64.0M   0% /dev
tmpfs                     1.9G         0      1.9G   0% /sys/fs/cgroup
/dev/sdj                246.0G     59.1M    233.4G   0% /bk/journal
/dev/sdi                246.0G     59.1M    233.4G   0% /bk/ledgers

[adrianmo]

In order to resize PVCs in a Pravega cluster managed by the operator, we need to:

1. Make sure the underlying block storage provider supports volume resizing.
2. Create the operator logic to identify when a user is attempting to resize a PVC.
3. Wait until the PVC condition to changes to FileSystemResizePending
4. Restart all affected pods for the PVC resizing to be applied (maybe sequentially to avoid downtime)
5. Check that the PVC size has changed

Related links:

• https://kubernetes.io/blog/2018/07/12/resizing-persistent-volumes-using-kubernetes/
• StatefulSet: support resize pvc storage in K8s v1.11 kubernetes/kubernetes#68737

[pbelgundi]

Moved to Bookkeeper-Operator: pravega/bookkeeper-operator#40