Check for pvc populated before doing any clone validations and actions #2375

ShellyKa13 · 2022-07-26T16:13:35Z

In case our pvc is already populated no need to check for source PVC
unknown size etc.

Missing tests

What this PR does / why we need it:

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Fixes #
This should fix bug: https://bugzilla.redhat.com/show_bug.cgi?id=2104479

Special notes for your reviewer:

Release note:

Fix clone datavolume with populated target PVC without source

mhenriks · 2022-07-27T01:22:33Z

pkg/controller/datavolume-controller.go

@@ -808,6 +806,9 @@ func (r *DatavolumeReconciler) ensureExtendedToken(pvc *corev1.PersistentVolumeC
 func (r *DatavolumeReconciler) selectCloneStrategy(datavolume *cdiv1.DataVolume, pvcSpec *corev1.PersistentVolumeClaimSpec) (cloneStrategy, error) {
 	preferredCloneStrategy, err := r.getCloneStrategy(datavolume)
 	if err != nil {
+		if k8serrors.IsNotFound(err) {
+			return NoClone, nil


maybe it would be better if getCloneStrategy() handles this case?

what should be returned in that case? preferredCloneStrategy=nil and err=nil? I would still need to add here a check if preferredCloneStrategy=nil and return NoClone. or create a new CDICloneStrategy-> cdiv1.CloneStrategyNoClone? I would still need to add here a check of if preferredCloneStrategy=cdiv1.CloneStrategyNoClone and return NoClone?
I dont mind doing either but doesnt seem any cleaner or simpler then this

Anyway, why do we want to ignore notFound Err?
For example if source is not found it can return notFoundError, and for CDI it means that it needs to exit and retry later. And this is what happens when we return NoClone, err

f err != nil { return reconcile.Result{}, err }

And how it behaves If we return NoClone, nil

If this behaves correctly then I am ok with the rest.

@brybacki I think it behaves as expected: If the source PVC doesn't exist but the target is already populated, we are just updating the DV without triggering any errors. However, if the source PVC doesn't exist and the target is not populated, it'll be handled by validateCloneAndSourcePVC, which will exit without errors to retry once the source is created:

containerized-data-importer/pkg/controller/datavolume-controller.go

Lines 673 to 682 in 083ee01

if pvcPopulated || prePopulated {

return r.reconcileDataVolumeStatus(datavolume, pvc, selectedCloneStrategy)

}

// Check if source PVC exists and do proper validation before attempting to clone

if done, err := r.validateCloneAndSourcePVC(datavolume); err != nil {

return reconcile.Result{}, err

} else if !done {

return reconcile.Result{}, nil

}

alromeros

Looks good!

alromeros · 2022-07-28T11:01:37Z

pkg/controller/datavolume-controller.go

@@ -2836,6 +2836,7 @@ func (r *DatavolumeReconciler) validateCloneAndSourcePVC(datavolume *cdiv1.DataV
 	if err != nil {
 		// Clone without source
 		if k8serrors.IsNotFound(err) {
+			r.recorder.Eventf(datavolume, corev1.EventTypeWarning, ErrUnableToClone, "Source pvc %s not found", datavolume.Spec.Source.PVC.Name)


Overall looks very good, but wouldn't this event be a little bit redundant? We are just recording a very similar event in the line below

I just moved it from the get clone strategy to here since it didnt seem related there, you added the event so you proabably know best if its redundant, if it is Ill delete :)

The first event says "Source pvc %s not found" while the second one is "The source PVC doesn't exist", so I think they are too similar to record them at the same time. Maybe it's a good idea to merge the two so the second event also displays the PVC name, but I'd just use one.

alromeros · 2022-07-28T11:31:29Z

pkg/controller/datavolume-controller.go

@@ -1904,7 +1905,6 @@ func (r *DatavolumeReconciler) getCloneStrategy(dataVolume *cdiv1.DataVolume) (*
 	sourcePvc, err := r.findSourcePvc(dataVolume)
 	if err != nil {
 		if k8serrors.IsNotFound(err) {


Just in case, you left an empty condition here.

right thanks :)

alromeros · 2022-07-29T08:42:37Z

/test pull-containerized-data-importer-e2e-hpp-latest

mhenriks · 2022-08-02T01:22:57Z

tests/cloner_test.go

+					targetName,
+					"1Gi",
+					map[string]string{
+						controller.AnnPodPhase:     "Succeeded",


What happens if this annotation is missing?

checked it, the test still passes, should I remove this annotation from the test?

alromeros · 2022-08-05T10:23:54Z

/test pull-containerized-data-importer-e2e-ceph

awels · 2022-08-05T19:10:42Z

tests/cloner_test.go

+				_, err = utils.CreateDataVolumeFromDefinition(f.CdiClient, f.Namespace.Name, cloneDV)
+				Expect(err).ToNot(HaveOccurred())
+				By("Wait for clone DV Succeeded phase")
+				err = utils.WaitForDataVolumePhaseWithTimeout(f, f.Namespace.Name, cdiv1.Succeeded, targetName, cloneCompleteTimeout)


Should we check to make sure the clone didn't happen? Maybe check for some logging in the controller? Or maybe check if the annCloneType annotation is not set.

In case our pvc is already populated no need to check for source PVC unknown size etc. Signed-off-by: Shelly Kagan <skagan@redhat.com>

ShellyKa13 · 2022-08-08T14:28:19Z

/test pull-containerized-data-importer-e2e-ceph-gc

mhenriks · 2022-08-08T15:06:55Z

/lgtm
/approve

kubevirt-bot · 2022-08-08T15:07:05Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: mhenriks

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

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [mhenriks]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

ShellyKa13 · 2022-08-09T06:26:20Z

/retest

ShellyKa13 · 2022-08-09T13:46:06Z

/retest

ShellyKa13 · 2022-08-09T14:36:40Z

/hold
checking if it causes a test flakiness

ShellyKa13 · 2022-08-10T08:32:56Z

/unhold
the test flake I saw is not because of my changes:
[test_id:6612]Clone with CSI as PVC source with target name that already exists
failed on timeout to get the clone type annotation on the dv
It will fail in case the target pvc is created before the data volume is being reconciled at least once,
in that case we will fail on the check of the pvc already exists before we will get to the clone reconcile which there we update the clone type on the datavolume annotation

ShellyKa13 · 2022-08-10T08:33:06Z

/retest

ShellyKa13 · 2022-08-10T09:07:10Z

opened a fix for the test flake: #2386

ShellyKa13 · 2022-08-17T13:45:35Z

/cherrypick release-v1.53

kubevirt-bot · 2022-08-17T13:46:12Z

@ShellyKa13: #2375 failed to apply on top of branch "release-v1.53":

Applying: Check for pvc populated before doing any clone validations and actions
Using index info to reconstruct a base tree...
M	pkg/controller/datavolume-controller.go
M	tests/cloner_test.go
Falling back to patching base and 3-way merge...
Auto-merging tests/cloner_test.go
Auto-merging pkg/controller/datavolume-controller.go
CONFLICT (content): Merge conflict in pkg/controller/datavolume-controller.go
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0001 Check for pvc populated before doing any clone validations and actions
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

In response to this:

/cherrypick release-v1.53

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/test-infra repository.

kubevirt#2375) In case our pvc is already populated no need to check for source PVC unknown size etc. Signed-off-by: Shelly Kagan <skagan@redhat.com> Signed-off-by: Shelly Kagan <skagan@redhat.com>

* Status reporting for CSI & Smart clones with WFFC storage (#2364) * Fix logging level so we respect it in controllers/operator Signed-off-by: Alex Kalenyuk <akalenyu@redhat.com> * Fix CSI & Smart clones with WFFC storage status reporting Signed-off-by: Alex Kalenyuk <akalenyu@redhat.com> * Check for pvc populated before doing any clone validations and actions (#2375) In case our pvc is already populated no need to check for source PVC unknown size etc. Signed-off-by: Shelly Kagan <skagan@redhat.com> Signed-off-by: Shelly Kagan <skagan@redhat.com> Signed-off-by: Alex Kalenyuk <akalenyu@redhat.com> Signed-off-by: Shelly Kagan <skagan@redhat.com> Co-authored-by: Shelly Kagan <78472213+ShellyKa13@users.noreply.github.com>

kubevirt-bot requested review from aglitke and awels July 26, 2022 16:28

mhenriks reviewed Jul 27, 2022

View reviewed changes

ShellyKa13 force-pushed the fix-clone-populated branch from ffac790 to 4ed66bc Compare July 28, 2022 08:30

ShellyKa13 changed the title ~~[wip] Check for pvc populated before doing any clone validations and actions~~ Check for pvc populated before doing any clone validations and actions Jul 28, 2022

kubevirt-bot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Jul 28, 2022

ShellyKa13 requested a review from mhenriks July 28, 2022 08:31

alromeros reviewed Jul 28, 2022

View reviewed changes

ShellyKa13 force-pushed the fix-clone-populated branch from 4ed66bc to cc119fa Compare July 28, 2022 16:13

ShellyKa13 requested a review from alromeros July 28, 2022 16:13

mhenriks reviewed Aug 2, 2022

View reviewed changes

kubevirt-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 3, 2022

ShellyKa13 force-pushed the fix-clone-populated branch from cc119fa to 7dd01a0 Compare August 4, 2022 14:25

kubevirt-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 4, 2022

ShellyKa13 force-pushed the fix-clone-populated branch from 7dd01a0 to 083ee01 Compare August 4, 2022 14:33

awels reviewed Aug 5, 2022

View reviewed changes

kubevirt-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 5, 2022

ShellyKa13 force-pushed the fix-clone-populated branch from 083ee01 to e0e6ef9 Compare August 7, 2022 12:06

kubevirt-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 7, 2022

Check for pvc populated before doing any clone validations and actions

d3ed3b9

In case our pvc is already populated no need to check for source PVC unknown size etc. Signed-off-by: Shelly Kagan <skagan@redhat.com>

ShellyKa13 force-pushed the fix-clone-populated branch from e0e6ef9 to d3ed3b9 Compare August 8, 2022 07:59

kubevirt-bot assigned mhenriks Aug 8, 2022

kubevirt-bot added the lgtm Indicates that a PR is ready to be merged. label Aug 8, 2022

kubevirt-bot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Aug 8, 2022

kubevirt-bot added the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Aug 9, 2022

kubevirt-bot removed the do-not-merge/hold Indicates that a PR should not merge because someone has issued a /hold command. label Aug 10, 2022

kubevirt-bot merged commit 5364b4b into kubevirt:main Aug 10, 2022

This was referenced Aug 22, 2022

[release-v1.43] Manual fix for clone without source pvc but target pvc already populated #2405

Merged

[release-v1.49] Manual fix for clone without source pvc but target pvc already populated #2406

Merged

brybacki mentioned this pull request Aug 26, 2022

[release-v1.43] Redo 2405 #2412

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Check for pvc populated before doing any clone validations and actions #2375

Check for pvc populated before doing any clone validations and actions #2375

ShellyKa13 commented Jul 26, 2022 •

edited

Loading

mhenriks Jul 27, 2022

ShellyKa13 Jul 27, 2022

brybacki Aug 4, 2022

brybacki Aug 4, 2022

alromeros Aug 5, 2022

brybacki Aug 8, 2022

alromeros left a comment

alromeros Jul 28, 2022

ShellyKa13 Jul 28, 2022

alromeros Jul 28, 2022

ShellyKa13 Jul 28, 2022

alromeros Jul 28, 2022

ShellyKa13 Jul 28, 2022

alromeros commented Jul 29, 2022

mhenriks Aug 2, 2022

ShellyKa13 Aug 4, 2022

ShellyKa13 Aug 4, 2022

alromeros commented Aug 5, 2022

awels Aug 5, 2022

ShellyKa13 commented Aug 8, 2022

mhenriks commented Aug 8, 2022

kubevirt-bot commented Aug 8, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 17, 2022

kubevirt-bot commented Aug 17, 2022

	if pvcPopulated \|\| prePopulated {
	return r.reconcileDataVolumeStatus(datavolume, pvc, selectedCloneStrategy)
	}

	// Check if source PVC exists and do proper validation before attempting to clone
	if done, err := r.validateCloneAndSourcePVC(datavolume); err != nil {
	return reconcile.Result{}, err
	} else if !done {
	return reconcile.Result{}, nil
	}

Check for pvc populated before doing any clone validations and actions #2375

Check for pvc populated before doing any clone validations and actions #2375

Conversation

ShellyKa13 commented Jul 26, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alromeros left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alromeros commented Jul 29, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alromeros commented Aug 5, 2022

Choose a reason for hiding this comment

ShellyKa13 commented Aug 8, 2022

mhenriks commented Aug 8, 2022

kubevirt-bot commented Aug 8, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 9, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 10, 2022

ShellyKa13 commented Aug 17, 2022

kubevirt-bot commented Aug 17, 2022

ShellyKa13 commented Jul 26, 2022 •

edited

Loading