daemon: don't go ready until CNI configuration has been written #32168
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
squeed
added
kind/bug
If the daemon is configured to write a CNI configuration file, we should not go ready until that CNI configuration file has been written. This prevents a race condition where the controller removes the taint from a node too early, meaning pods may be created with a different CNI provider. In cilium#29405, Cilium was configured in chaining mode, but the "primary" CNI provider hadn't written its configuration yet. This caused the not-ready taint to be removed from the node too early, and pods were created in a bad state. By hooking in the CNI cell's status in the daemon's Status type, we prevent the daemon's healthz endpoint from returning a successful response until the CNI cell has been successful. Fixes: cilium#29405 Signed-off-by: Casey Callendrello <cdc@isovalent.com>
squeed
force-pushed
the
cni-chained-readiness
branch
from
f03e628
to
ec776ff
Compare
squeed
added
the
needs-backport/1.15
|
(removed some api changes I realized I didnt' want) |
|
/test |
maintainer-s-little-helper
bot
added
the
ready-to-merge
|
Awesome! Will this be backported to 1.14 too? |
|
1.15 for sure, I’m not sure about 1.14 yet |
squeed
added
the
needs-backport/1.14
|
Looks like this applies mostly cleanly to v1.14 too; marked as such. |
gandro
added
backport-pending/1.15
gandro
added
backport-pending/1.14
github-actions
bot
added
backport-done/1.14
This was referenced May 10, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
If the daemon is configured to write a CNI configuration file, we should not go ready until that CNI configuration file has been written. This prevents a race condition where the controller removes the taint from a node too early, meaning pods may be created with a different CNI provider.
In #29405, Cilium was configured in chaining mode, but the "primary" CNI provider hadn't written its configuration yet. This caused the not-ready taint to be removed from the node too early, and pods were created in a bad state.
By hooking in the CNI cell's status in the daemon's Status type, we prevent the daemon's healthz endpoint from returning a successful response until the CNI cell has been successful.
Fixes: #29405