Bug 1959238 - CVO creating cloud-controller-manager too early causing upgrade failures
Summary: CVO creating cloud-controller-manager too early causing upgrade failures
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Cluster Version Operator
Version: 4.8
Hardware: Unspecified
OS: Unspecified
high
medium
Target Milestone: ---
: 4.7.z
Assignee: W. Trevor King
QA Contact: Yang Yang
URL:
Whiteboard:
Depends On: 1957775
Blocks: 1943566
TreeView+ depends on / blocked
 
Reported: 2021-05-11 04:45 UTC by OpenShift BugZilla Robot
Modified: 2021-06-01 04:51 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Cause: The cluster-version operator had been pre-creating ClusterOperator resources at the beginning of the update phase. Consequence: For new second-level operators who were deeper in in the update, that could leave their ClusterOperator without status conditions for many minutes. After ten minutes without conditions, the ClusterOperatorDown and/or ClusterOperatorDegraded alerts would begin firing. Fix: Now the cluster-version operator pre-creates ClusterOperator resources much more closely to the creation of the new second-level operator's deployment. Result: The new second-level operator has time to come up and set status.conditions on the ClusterOperator before alarms start firing.
Clone Of:
Environment:
Last Closed: 2021-06-01 04:50:27 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github openshift cluster-version-operator pull 557 0 None open Bug 1959238: pkg/cvo/sync_worker: Shift ClusterOperator pre-creation into the manifest-task node 2021-05-11 06:12:20 UTC
Red Hat Product Errata RHSA-2021:2121 0 None None None 2021-06-01 04:51:39 UTC

Comment 2 Yang Yang 2021-05-25 06:04:06 UTC
CCM is not included in nightly build. So verifying with ci build 4.7.0-0.ci-2021-05-22-053635.

# oc get co | grep cloud-controller-manager
null

cloud-controller-manager is not included in 4.7 ci build

Upgrade to 4.8 ci build

# oc get clusterversion
NAME      VERSION                        AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.7.0-0.ci-2021-05-22-053635   True        True          10m     Working towards 4.8.0-0.ci-2021-05-24-203349: 93 of 694 done (13% complete)

Watching the cloud-controller-manager,

# oc get co -w | grep cloud-controller-manager

cloud-controller-manager                                                                                       
cloud-controller-manager                                                                                       
cloud-controller-manager                   4.8.0-0.ci-2021-05-24-203349   True        False         False      0s

CVO creates cloud-controller-manager and soon condition status is visible.

# oc get clusterversion
NAME      VERSION                        AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.8.0-0.ci-2021-05-24-203349   True        False         60m     Cluster version is 4.8.0-0.ci-2021-05-24-203349

There are no alert ClusterOperatorDown and/or ClusterOperatorDegraded for cloud-controller-manager fired. And upgrade is successful.

The pr works. Pending regression test against nightly build once the build is available.

Comment 4 Yang Yang 2021-05-26 09:33:18 UTC
Performing regression test against 4.7.0-0.nightly-2021-05-25-192356 by upgrading to 4.8.0-0.nightly-2021-05-25-223219.

# oc get clusterversion
NAME      VERSION                             AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.7.0-0.nightly-2021-05-25-192356   True        True          44m     Working towards 4.8.0-0.nightly-2021-05-25-223219: 549 of 675 done (81% complete), waiting up to 40 minutes on dns

# oc get clusterversion
NAME      VERSION                             AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.8.0-0.nightly-2021-05-25-223219   True        False         5m24s   Cluster version is 4.8.0-0.nightly-2021-05-25-223219

There are no alert ClusterOperatorDown and/or ClusterOperatorDegraded fired. And upgrade is successful.

Moving it to verified state.

Comment 6 errata-xmlrpc 2021-06-01 04:50:27 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: OpenShift Container Platform 4.7.13 bug fix and security update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2021:2121


Note You need to log in before you can comment on or make changes to this bug.