-
Notifications
You must be signed in to change notification settings - Fork 532
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove two types of nacks - "Nonexistent client" & "Readonly client" #7753
Merged
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
(cherry picked from commit 719c03f131d025cd8f74db74e1a716dedab69911)
jatgarg
approved these changes
Oct 6, 2021
jatgarg
reviewed
Oct 8, 2021
jatgarg
reviewed
Oct 8, 2021
vladsud
added a commit
that referenced
this pull request
Oct 8, 2021
…DDS tests (#7784) Slight changes in reconnect logic (see PRs #7753, #7393) result in failures in PropertyDDS UTs. Looking a bit deeper, I can't easily follow intentions in OpProcessingController. For example, OpProcessingController.process(dm1) will leave all but dm1 paused. UTs do rely on that (it's pretty clearly from some of them), in other places it feels like it's unintended result. And I do not think it correctly waits for all pending activity to be flushed (as observed in above mentioned PRs). Given that it's deprecated and repo uses LoaderContainerTracker (explicitly or implicitly through), it's time to do the move. Some random tests fail with this change, but are addressed by increasing timeout. We saw this problem before with these tests - see pending #7350
github-actions
bot
added
the
area: tests
Tests to add, test infrastructure improvements, etc
label
Oct 9, 2021
⯅ @fluidframework/base-host: +1.21 KB
■ @fluid-example/bundle-size-tests: No change
Baseline commit: 5fc2cbf |
This was referenced Dec 6, 2021
vladsud
added a commit
that referenced
this pull request
Dec 7, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
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.
This is part of looking into #7137 (nacks, specifically - tight nack loops).
This change adds logic for DeltaManager to realize when it does not have "write" connection and stop sending ops on such connections. This removes "expected" nacks from the workflow. Remaining nack types are (in most cases) logical errors that needs to be raised in telemetry and looked into (10K unsummarized ops, 1Mb+ ops, etc.).
Removing "expected" nacks from the workflows allows us to make next step - apply some kind of workflow that breaks tight infinite loop of remaining nacks types. For example, we may add delays, or better - close container if we see number of nacks in a row (this is not part of this change, I'll continue to use this issue to track this work).
The con of this approach - it adds more logic to already complicated DeltaManager.
One of the other pros - it makes things more consistent to other layers. I.e. if Container checks for connection mode and it's "write", it knows for sure that it can send an op. Even more important - client knows that it should not send ops (for example, consensus / leader election DDSs would stop sending ops, and allow this client to stay in "read" connection mode once it got there).