-
Spike
-
Resolution: Done
-
Normal
-
None
-
None
-
None
-
5
-
False
-
None
-
False
-
BIFROST-105 - Experience #3: OCP Node Customization
-
-
-
MCO Sprint 244, MCO Sprint 245, MCO Sprint 246, MCO Sprint 247, MCO Sprint 248
-
0
-
0.000
Currently, BuildController looks at the machine-config-osimageurl ConfigMap in the MCO namespace to determine which base image to use. Whenever an OpenShift upgrade occurs, that value could change and if the OS image is not rebuilt, there could be a Kube version mismatch, amongst other maladies. As a result, BuildController should rebuild the OS image. Right now, it likely will not do that.
Done When:
- We have determined if on-cluster builds works as it should during an OpenShift upgrade. In other words, that a rebuild occurs when the new OS image is available.
- If the above case is not met, then additional stories will need to be opened.
- Determine what happens when an osImageURL override occurs although that may be better covered by https://issues.redhat.com/browse/MCO-696.
- is related to
-
MCO-1000 Decouple the image build and image rollout processes
-
- To Do
-
-
-
- To Do
-
-
-
- In Progress
-
-
-
- In Progress
-
-
-
- In Progress
-
-
OCPBUGS-24705 Opting into on-cluster builds sometimes does not respect maxUnavailable setting
-
- Closed
-