Refactor AsyncUploader to replace Engine.Unit with ComponentManager #6809
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
Replace engine.Unit with component.Component - Arbitrarily set number of workers to 3 Uses a fifoqueue to buffer tasks (planned to replace with channel, since we would rather wait to upload than let a block be not uploaded). Does not correctly propagate errors, or calculate any new metrics yet.
BadgerRetryableUploadWrapper wraps an AsyncUploader. Now that the AsyncUploader is a Component, it needs to be `Start()`ed. The retryable wrapper did not itself do anything special on ready/done, so the Component functionality is directly delegated to the wrapped AsyncUploader.
Since AsyncUploader now implements Component instead of ReadyDoneAware, update the methods used to start and end the AsyncUploader (using a context and its cancel() function). Test "uploads are run in parallel" currently failing (due to only one worker taking tasks from the queue?)
Instead of using a notifier and fifoqueue with metrics, use a buffered channel. Block when the channel is full. (Reasoning: we want to ensure no uploads get dropped.) Buffer size of 100 was chosen arbitrarily. All AsyncUploader tests now pass.
Because Component's `Ready()` and `Done()` methods work differently from ReadyDoneAware's, include them in Component with Component-specific comments. Namely, specify that Components should be started with Start() and shutdown by canceling the context they were started with.
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## feature/pusher-engine-refactor #6809 +/- ##
==================================================================
- Coverage 41.25% 41.24% -0.01%
==================================================================
Files 2062 2062
Lines 182730 182742 +12
==================================================================
- Hits 75383 75375 -8
- Misses 101033 101051 +18
- Partials 6314 6316 +2
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
|
One thing I haven't done anything about yet - we might want to rewrite the AsyncUploader tests to make more use of testing utility functions for ensuring synchronization / counting function calls, rather than explicitly use waitgroups all the time. Something to consider, though the synchronization behaviour of some of the current tests seems a bit complex. |
Update/clarify doc comments Co-authored-by: Jordan Schalm <[email protected]>
zhangchiqing
approved these changes
Dec 18, 2024
peterargue
reviewed
Dec 19, 2024
Co-authored-by: Jordan Schalm <[email protected]>
Channel size of 20000 and worker count of 100 suggested by Leo. 20000 is approximately equal to 4 hours of execution results. Co-authored-by: Leo Zhang <[email protected]>
jordanschalm
approved these changes
Dec 20, 2024
…ader-engine.Unit-refactor
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.
AsyncUploader now uses the Component interface, and a channel is used to buffer upload tasks until a worker goroutine is available to begin the upload.
In addition, updated documentation comments for the Component interface to clarify differences from ReadyDoneAware.
Part of #6807.