-
Story
-
Resolution: Done
-
Normal
-
None
-
None
Description
This story covers
- ensuring the WICD controller service is not running before running WICD cleanup. This is needed to recover from a node config failure that occurs after the WICD service is started
- using the new version's WICD executable to cleanup Windows services on the node during upgrade. This is important so that bug fixes in WICD are picked up and used during the upgrade flow
Engineering Details
- copy over and use the new WMCO version's WICD binary to cleanup during upgrade
- stop the WICD controller service from running before running WICD cleanup when recovering from node config failure
Acceptance Criteria
During upgrade from WMCO version A to version B:
- The WICD Windows service is not running when the WICD cleanup command is invoked
- WICD executable from version B is executed to cleanup node services
Testing
QE: This will require a full upgrade test, upgrading cluster OCP version and WMCO major version. Please test
- upgrade from 8.0.0 to 9.0.0 when this merges into master
- from 7.1.0 to 8.y.z when the backport lands in release-4.13
- from 6.0.0 to 7.y.z when the backport lands in release-4.12
- links to
-
RHSA-2023:119726 Red Hat OpenShift support for Windows Containers 7.2.0 security update
- mentioned on
(1 links to, 3 mentioned on)