Bug 1979672 - Prevent cluster upgrade if MCP are paused
Summary: Prevent cluster upgrade if MCP are paused
Keywords:
Status: CLOSED DEFERRED
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Machine Config Operator
Version: 4.6
Hardware: Unspecified
OS: Unspecified
unspecified
medium
Target Milestone: ---
: ---
Assignee: Sinny Kumari
QA Contact: Michael Nguyen
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-07-06 17:05 UTC by oarribas
Modified: 2024-12-20 20:24 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-07-09 14:01:29 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker RFE-1993 0 High Backlog Not allow to start a cluster upgrade if master MCP is paused 2021-07-09 14:01:29 UTC
Red Hat Knowledge Base (Solution) 5317441 0 None None None 2021-07-06 17:07:25 UTC
Red Hat Knowledge Base (Solution) 5477811 0 None None None 2021-07-06 17:07:25 UTC

Description oarribas 2021-07-06 17:05:42 UTC
Description of problem:

If MCP are paused while upgrading the cluster, the upgrade is not finishing and the machine-config cluster operator becomes degraded.


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

All OCP 4.x


How reproducible:

Always


Steps to Reproduce:
1. Pause the MCP [1]
2. Upgrade the cluster
3.

Actual results:

machine-config degraded [2].


Expected results:

A message should be shown when trying to upgrade the cluster, and the upgrade should not start while the MCP are paused.


Additional info:

Could be use cases where the upgrade should be executed even if MCP paused. Maybe an option to not check the paused MCP could be added for those cases.



[1] https://access.redhat.com/solutions/5477811
[2] https://access.redhat.com/solutions/5317441

Comment 1 Yu Qi Zhang 2021-07-06 17:26:06 UTC
> If MCP are paused while upgrading the cluster, the upgrade is not finishing and the machine-config cluster operator becomes degraded.

Do you have a must-gather for this?

Generally speaking, we expect to allow upgrades if the worker MCP is paused. This is the mechanism we plan to use to skip versions when upgrading. Or are you talking about masters?


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