[v1.15] fix: Assign PodStore from Pod resource until cell migration is completed #35419
Conversation
[ upstream commit c3db8a5 ] Fixes cilium#32713 PodNodeNameIndex is added to the Pod watcher in operatorK8s.Resource. This will be refactored once `ciliumNodeSynchronizer` is migrated to a cell. Signed-off-by: Dorde Lapcevic <dordel@google.com> Signed-off-by: Anton Ippolitov <anton.ippolitov@datadoghq.com>
5d1fe6d to
a90d47b
Compare
|
Thank you @antonipp! |
|
/test-backport-1.15 |
|
@dlapcevic @antonipp FYI, #34090 (comment). Could we fix this before we go and release this patch in another version without a proper release note? |
|
Looks good to me. I think we can move forward with merging it. @antonipp please confirm. |
|
I added a release note to my PR to indicate that a bug was fixed. I don't have rights to change the labels on the PR though. |
I believe the release note and label need to be on the upstream PR. I'm unsure what effect they would have on the backport PR. AFAIK, we derive release notes for backport PRs from the upstream PR. |
|
Ok, I didn't know it worked this way, I will let Dorde update the original PR then. |
|
cc @dlapcevic |
Backport #34090 to 1.15
A couple of notes:
PodResourcetooperator/k8s/resource_ctors.gowhich was created recently in pkg/k8s: Add required resources for Operator managing CIDs #33021 and didn't exist in 1.15I tested the backport in our cluster running 1.15 and confirmed that "Unable to compute pending pods, will not surge-allocate" warnings were gone.