Bug 2313492: [release-4.14] Remove check of object type & return in enqueueStorageClusterRequest #2811
Conversation
The enqueueStorageClusterRequest event handler is used to enqueue a reconcile request for the controller from the watched objects. Although the handle is being used by all the watched objects, but it checks the object type and returns if the object is not of type OCSInitialization. Due to this when any thing happens with the watched objects other than OCSInitialization, the reconcile request is not enqueued and the controller does not reconcile the object. Signed-off-by: Malay Kumar Parida <mparida@redhat.com>
openshift-ci
bot
added
bugzilla/severity-high
|
@malayparida2000: This pull request references Bugzilla bug 2313492, which is invalid:
Comment In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: malayparida2000 The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
@malayparida2000: This pull request references Bugzilla bug 2313492, which is invalid:
Comment In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
/retest |
|
@malayparida2000: The following test failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here. |
The enqueueStorageClusterRequest event handler is used to enqueue a reconcile request for the controller from the watched objects. Although the handle is being used by all the watched objects, but it checks the object type and returns if the object is not of type OCSInitialization. Due to this when any thing happens with the watched objects other than OCSInitialization, the reconcile request is not enqueued and the controller does not reconcile the object.
The watch process on certain resources we had on our codebis broken in 4.14. But the issue was being masked due to the side effect of other problem where we were doing continuous reconcile each second.Recently Nitin fixed that Bug so now unnecessary reconciles don't happen anymore,#2624.
So, Now the issue was noticed as creation of the Virtual Machine CRD was not triggering a reconcile and thus not creating the special Storageclass that it should have. Upon debugging we found the enqueue function we use on 4.14 just returns if the resource is not OCSInitialization, but we are still using that enqueue function for all the watches we have. Hence a reconcile request never gets queued for the Watched Resources.
The issue is not present on 4.15 onwards as I had unintentionally fixed the bug while working on the Separation of ceph toolbox from ocsinitialization as a part of the Multiple Storagecluster Epic in 4.15, #2240.
Hence this bug requires specific fix only in 4.14.z.