Bug 2227761 - openshift-monitoring apiservice on FailedDiscoveryCheck
Summary: openshift-monitoring apiservice on FailedDiscoveryCheck
Keywords:
Status: ASSIGNED
Alias: None
Product: Container Native Virtualization (CNV)
Classification: Red Hat
Component: Metrics
Version: 4.14.0
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: 4.14.0
Assignee: Shirly Radco
QA Contact: Natalie Gavrielov
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2023-07-31 10:17 UTC by Roni Kishner
Modified: 2023-08-08 12:40 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed:
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker CNV-31552 0 None None None 2023-07-31 10:17:49 UTC

Description Roni Kishner 2023-07-31 10:17:27 UTC
Description of problem:
After trying to delete a namespace with some resource inside (Secret/VM), the apiservice resource 'openshift-monitoring/prometheus-adapter' is on FailedDiscoveryCheck, this causes the namespace deletion to be stuck on 'Terminating'

After deleting the apiservice (oc delete apiservice v1beta1.metrics.k8s.io) the namespace do get deleted after 1-3 minutes, and the issue does not repeat.

Version-Release number of selected component (if applicable):
4.14

How reproducible:
Need more investigation but seen more then a few times on different clusters

Steps to Reproduce:
1. Create a VM in a namespace
2. Delete the namespace

Actual results:
The namespace is stuck on Terminating

Expected results:
The namespace and all resources should be deleted

Additional info:
It is not clear if the apiservice failure is a side effect of another issue or it is what is causing this.

Comment 1 Kedar Bidarkar 2023-08-02 12:35:03 UTC
Changing the component to Metrics, so that your team  could triage this bug. 
Discussed with Virt Devs and feel it could be related to Metrics thing.


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