lib/resourcebuilder/apps: Only error on Deployment Available=False *and* Progressing=False #430
+3
−12
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…nd* Progressing=False Available=True, Progressing=False is the happy, steady state. Available=True, Progressing=True is a happy update. Available=False, Progressing=True is acceptable outage, e.g. during an update with the Recreate strategy [1]: $ curl -s https://storage.googleapis.com/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade-4.6/1291426211527921664/artifacts/e2e-gcp-upgrade/container-logs/test.log | grep MinimumReplicasUnavailable | head -n1 Aug 6 17:56:00.674: INFO: deployment status: v1.DeploymentStatus{ObservedGeneration:1, Replicas:1, UpdatedReplicas:1, ReadyReplicas:0, AvailableReplicas:0, UnavailableReplicas:1, Conditions:[]v1.DeploymentCondition{v1.DeploymentCondition{Type:"Available", Status:"False", LastUpdateTime:v1.Time{Time:time.Time{wall:0x0, ext:63732333358, loc:(*time.Location)(0x9e74040)}}, LastTransitionTime:v1.Time{Time:time.Time{wall:0x0, ext:63732333358, loc:(*time.Location)(0x9e74040)}}, Reason:"MinimumReplicasUnavailable", Message:"Deployment does not have minimum availability."}, v1.DeploymentCondition{Type:"Progressing", Status:"True", LastUpdateTime:v1.Time{Time:time.Time{wall:0x0, ext:63732333358, loc:(*time.Location)(0x9e74040)}}, LastTransitionTime:v1.Time{Time:time.Time{wall:0x0, ext:63732333358, loc:(*time.Location)(0x9e74040)}}, Reason:"ReplicaSetUpdated", Message:"ReplicaSet \"dp-7f9df745ff\" is progressing."}}, CollisionCount:(*int32)(nil)} Available=False, Progressing=False is the Deployment controller saying "I cannot deliver my expected service level for this Deployment", so that's when we should be complaining. Fixes noise like: Aug 6 18:03:00.500: INFO: cluster upgrade is Failing: Multiple errors are preventing progress: * Could not update namespace "openshift-service-ca-operator" (467 of 608) * deployment openshift-cluster-machine-approver/machine-approver is not available MinimumReplicasUnavailable: Deployment does not have minimum availability. * deployment openshift-ingress-operator/ingress-operator is not available MinimumReplicasUnavailable: Deployment does not have minimum availability. (the namespace part of that message is a separate issue). [1]: https://prow.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade-4.6/1291426211527921664
openshift-ci-robot
added
the
approved
Contributor
|
I think this looks right, I'm going to trigger an upgrade via clutser bot between 4.5 and this. |
Contributor
Contributor
|
Logs in that test run look correct |
Contributor
|
/lgtm |
Contributor
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: smarterclayton, wking The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Contributor
|
/retest |
Member
Author
|
In this space is the upstream kubernetes/kubernetes#93933 fix. In the meantime, the CVO will continue to assume that Deployments which fall into that hole are happy enough, even if they should have been set I think this PR is orthogonal enough that it should not block on any of the above paragraph getting addressed. |
Contributor
|
Agree, we should carry the fix to the workload controller for error detection, not CVO |
wking
added a commit
to wking/cluster-version-operator
that referenced
this pull request
Apr 20, 2021
See in a customer cluster [1]:
$ yaml2json <namespaces/openshift-marketplace/apps/deployments.yaml | jq -r '.items[].status.conditions'
[
{
"lastTransitionTime": "2021-04-12T22:04:41Z",
"lastUpdateTime": "2021-04-12T22:04:41Z",
"message": "Deployment has minimum availability.",
"reason": "MinimumReplicasAvailable",
"status": "True",
"type": "Available"
},
{
"lastTransitionTime": "2021-04-19T17:33:30Z",
"lastUpdateTime": "2021-04-19T17:33:30Z",
"message": "ReplicaSet \"marketplace-operator-f7cc88d59\" has timed out progressing.",
"reason": "ProgressDeadlineExceeded",
"status": "False",
"type": "Progressing"
}
]
Previous touch for this logic was 0442094
(lib/resourcebuilder/apps: Only error on Deployment Available=False
*and* Progressing=False, 2020-08-06, openshift#430), where I said:
Available=True, Progressing=False is the happy, steady state.
This commit tightens that back down a bit, to account for cases where
the Deployment controller claims the Deployment is still available,
but has given up on rolling out a requested update [2]. We definitely
want to complain about Deployments like that, because admins might
want to investigate the sticking issue, and possibly delete or
otherwise poke the Deployment so we come back in and reconcile it so
the Deployment controller gives it another attempt.
[1]: https://bugzilla.redhat.com/show_bug.cgi?id=1951339#c0
[2]: https://kubernetes.io/docs/concepts/workloads/controllers/deployment/#failed-deployment
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Available=True,Progressing=Falseis the happy, steady state.Available=True,Progressing=Trueis a happy update.Available=False,Progressing=Trueis acceptable outage, e.g. during an update with theRecreatestrategy:Available=False,Progressing=Falseis the Deployment controller saying "I cannot deliver my expected service level for this Deployment", so that's when we should be complaining. Fixes noise like:(the namespace part of that message is a separate issue).