fix: let NNCResonciler process update events with different generations when IPAMv2 is enabled #3279
+44
−27
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
Group the predicate funcs into either specific event filters or the "universal" filters which get applied to every event type. Then And() them which is what controller-runtime event_handler is functionally doing anyway. link: https://github.com/kubernetes-sigs/controller-runtime/blob/aea2e32a936584b06ae6f7992f856fe7292b0297/pkg/internal/source/event_handler.go#L116
timraymond
approved these changes
Dec 19, 2024
| @@ -1482,7 +1482,7 @@ func InitializeCRDState(ctx context.Context, httpRestService cns.HTTPService, cn | |||
| // wait for the Reconciler to run once on a NNC that was made for this Node. | |||
| // the nncReadyCtx has a timeout of 15 minutes, after which we will consider | |||
| // this false and the NNC Reconciler stuck/failed, log and retry. | |||
| nncReadyCtx, cancel := context.WithTimeout(ctx, 15*time.Minute) //nolint // it will time out and not leak | |||
| nncReadyCtx, cancel := context.WithTimeout(ctx, 15*time.Minute) // nolint // it will time out and not leak | |||
There was a problem hiding this comment.
I believe the nolint pragmas are intended to be directly next to the //
Suggested change
| nncReadyCtx, cancel := context.WithTimeout(ctx, 15*time.Minute) // nolint // it will time out and not leak | |
| nncReadyCtx, cancel := context.WithTimeout(ctx, 15*time.Minute) //nolint // it will time out and not leak |
There was a problem hiding this comment.
here's an arbitrary webpage to say that you're right. wonder why my auto formatter is changing these lines. https://golangci-lint.run/usage/false-positives/#:~:text=tests%20for%20it.-,Use,because,-machine%2Dreadable%20comments
will revert these.
There was a problem hiding this comment.
except I wonder if //nolint by itself does anything and you really need to do //nolint:all so my formatter thinks these are just normal comments ... ?
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.
Reason for Change:
IPAMv2 can handle update events even when the update is made from CNS itself, like patching
the spec for more IPs. This solves the problem of missed updates when a watch is re-established
and an NNC has been updated in both its spec and status. IPAMv2 is idempotent and will not
fall into a feedback loop where after CNS patches the NNC for more IPs its NNC reconciler still
sees IPs needed so it requests more IPs again, which triggers the reconciler again to request
more IPs, and so on ...
Issue Fixed:
Fixes #3278
Requirements:
Notes:
The first commit is the basic implementation. The 2nd commit does some more refactoring. I am fine to take just the first commit if you all don't like the refactoring attempt.