Revert "daemonset controller should respect taints" #31907
Conversation
mikedanese
added
the
release-note-none
davidopp
added
the
lgtm
|
LGTM |
|
Can one of the admins verify that this patch is reasonable to test? If so, please reply "ok to test". This message will repeat several times in short succession due to jenkinsci/ghprb-plugin#292. Sorry. |
1 similar comment
|
Can one of the admins verify that this patch is reasonable to test? If so, please reply "ok to test". This message will repeat several times in short succession due to jenkinsci/ghprb-plugin#292. Sorry. |
k8s-github-robot
added
the
size/M
|
@k8s-bot ok to test, issue: #IGNORE |
|
@k8s-bot node e2e test this please, issue: #IGNORE |
|
GCE e2e build/test passed for commit a615c42. |
|
Automatic merge from submit-queue |
|
When the time comes for this to be unreverted, please take a look at #31136 too. Also you may let me know what needs to be changed there so I can have it fixed by that time. Thanks! |
|
@mikedanese Why did you revert it? I don't see any bug for it. |
|
@lukaszo daemonset controller should not delete a daemon pod that is already running on a machine if the taint Effect on that node is NoSchedule. This is not the case with this change, we wanted to make sure this is correct when we introduce the behavior and don't want to slightly modify the behavior between v1.4 and v1.5 |
|
I see, thank you for the explanation. |
|
@mikedanese I just realized that the current 1.5 beta still schedules daemons onto nodes with NoSchedule taints. Wasn't it planned to unrevert this after 1.4? |
|
@codablock yes but I didn't get around to fixing the issue. I'm going to try very hard to get this into 1.6 at this point. |
Reverts #31020
We will be unreverting with some modifications after v1.4.
cc @pwittrock @davidopp
This change is