Bug 1925597 - kata-operator-daemon-uninstall pods did not exit after deleting kataconfig instance
Summary: kata-operator-daemon-uninstall pods did not exit after deleting kataconfig in...
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: sandboxed-containers
Version: 4.7
Hardware: Unspecified
OS: Unspecified
unspecified
medium
Target Milestone: ---
: ---
Assignee: Ariel Adam
QA Contact: Cameron Meadors
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-02-05 16:16 UTC by Cameron Meadors
Modified: 2021-09-16 20:10 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-09-16 20:10:38 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Cameron Meadors 2021-02-05 16:16:15 UTC
Description of problem:

When I deleted the kataconfig instance, the uninstall pods never exited leaving the system in an ambiguous state.

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

Installed kata-operator from the openshift/kata-operator on master.

How reproducible:

I can not reproduce it.

Steps to Reproduce:
1. Have a kataconfig instance created
2. oc delete kataconfig <instance>
3. 

Actual results:

uninstall pods never exited.  Kata runtime looked like it was indeed uninstalled from nodes.

Expected results:

Uninstall pods would exit after a short time (about 8 minutes) and runtime would be removed from nodes

Additional info:

I tried to get things to retrigger and clean up buy creating another kataconfig instance, but that resulted in the install pods working correctly and exiting, but the uninstall pods were still around.  This also resulted the kataconfig instance reporting that the runtime was both installed and being uninstalled on all the nodes. From 'oc describe kataconfig example-kataconfig:

Installation Status:
    Completed:
      Completed Nodes Count:  3
      Completed Nodes List:
        cmead-kata47-7-wcvj7-worker-a-9h7j9.c.openshift-qe.internal
        cmead-kata47-7-wcvj7-worker-b-x9grc.c.openshift-qe.internal
        cmead-kata47-7-wcvj7-worker-c-mqmwb.c.openshift-qe.internal
    Failed:
    In Progress:
  Kata Image:         
  Runtime Class:      kata
  Total Nodes Count:  3
  Un Installation Status:
    Completed:
    Failed:
    In Progress:
      Binaries Uninstall Nodes List:
        cmead-kata47-7-wcvj7-worker-a-9h7j9.c.openshift-qe.internal
        cmead-kata47-7-wcvj7-worker-b-x9grc.c.openshift-qe.internal
        cmead-kata47-7-wcvj7-worker-c-mqmwb.c.openshift-qe.internal
      In Progress Nodes Count:  3

After a few more cycles of installing and uninstalling the uninstall pods finally exited and nodes appeared to be in the correct state.  Uninstall pods had an error, but I unfortunately did not capture it.

Comment 3 Pradipta Banerjee 2021-09-16 09:36:44 UTC
Hi Cameron, can this bug be closed? It's no longer relevant with latest operator code using RHCOS extension.


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