ceph: initialize rbd block pool after creation

[Rakshith-R]

This is done in order to prevent deadlock when parallel
PVC create requests are issued on a new uninitialized
rbd block pool due to https://tracker.ceph.com/issues/52537.

Fixes: #8696

Signed-off-by: Rakshith R rar@redhat.com

verified by testing it locally

[rakshith@fedora ceph-csi]$ ./scripts/rook.sh create-block-pool
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   525  100   525    0     0   1299      0 --:--:-- --:--:-- --:--:--  1299
cephblockpool.ceph.rook.io/new-pool created
Checking RBD (new-pool) stats... 0s
RBD (new-pool) is successfully created...
[rakshith@fedora ceph-csi]$ k apply -f rbd/storageclass-test.yaml 
storageclass.storage.k8s.io/rook-ceph-block-1 created
[rakshith@fedora ceph-csi]$ k apply -f rbd/pvc.yaml 
persistentvolumeclaim/rbd-1 created
persistentvolumeclaim/rbd-2 created
persistentvolumeclaim/rbd-3 created
persistentvolumeclaim/rbd-4 created
[rakshith@fedora ceph-csi]$ k get pvc
NAME    STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS        AGE
rbd-1   Bound    pvc-81fbaf0e-aba8-4f18-8e4b-aeb18fb17527   1Gi        RWO            rook-ceph-block-1   34s
rbd-2   Bound    pvc-7e599340-1b14-42f7-9517-bb9733ded3f2   1Gi        RWO            rook-ceph-block-1   33s
rbd-3   Bound    pvc-ebb4d196-ee25-4369-b908-c91908324c9f   1Gi        RWO            rook-ceph-block-1   33s
rbd-4   Bound    pvc-e513679e-4c37-4236-84f0-ce0d044f87fe   1Gi        RWO            rook-ceph-block-1   33s

Checklist:

• Commit Message Formatting: Commit titles and messages follow guidelines in the developer guide.
• Skip Tests for Docs: Add the flag for skipping the build if this is only a documentation change. See here for the flag.
• Skip Unrelated Tests: Add a flag to run tests for a specific storage provider. See test options.
• Reviewed the developer guide on Submitting a Pull Request
• Documentation has been updated, if necessary.
• Unit tests have been added, if necessary.
• Integration tests have been added, if necessary.
• Pending release notes updated with breaking and/or notable changes, if necessary.
• Upgrade from previous release is tested and upgrade user guide is updated, if necessary.
• Code generation (make codegen) has been run to update object specifications, if necessary.

[leseb]

Is pool init idempotent?

[Rakshith-R]

AFAIK and according to my tests, pool init is idempotent.

I'll let @idryomov confirm this (and please review the pr too thanks).

[idryomov]

Yes, it is idempotent.

[Rakshith-R]

Moved it to higher layer, Since adding check for rbd through the function calls will make it complex.

@idryomov Shall i add a TODO to remove the cmd once we get the fix in cephcsi build or can it be considered a good practise to initialize the rbd blockpool after creation in rook?

[leseb]

It would be nice to add unit tests just like you did before.

[idryomov]

@idryomov Shall i add a TODO to remove the cmd once we get the fix in cephcsi build or can it be considered a good practise to initialize the rbd blockpool after creation in rook?

It is good to have this. This initialization can't be skipped -- it happens one way or the other anyway. With an explicit rbd pool init it is offloaded from the first image creation request, making all image creation requests the same and avoiding a couple of potential race conditions.

[idryomov]

LGTM from rbd perspective

[Rakshith-R]

It would be nice to add unit tests just like you did before.

Added them, PTAL.

[leseb]

One more thing, logging is important so let's go before and after the command has run, like initializing rbd pool %q and successfully initialized rbd pool %q. Thanks!