[SPARK-48314][SS] Don't double cache files for FileStreamSource using Trigger.AvailableNow #46627
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
|
@HeartSaVioR since you added the file caching originally back in the day |
Kimahriman
changed the title
HeartSaVioR
reviewed
May 20, 2024
| f | ||
| } | ||
|
|
||
| source.latestOffset(FileStreamSourceOffset(-1L), ReadLimit.maxFiles(5)) |
There was a problem hiding this comment.
Shall we check the result just for completeness sake?
There was a problem hiding this comment.
Updated to check files returned, and used the new setting from #45362 to verify correct number of files based on not caching
|
|
||
| // Reading again leverages the files already tracked in allFilesForTriggerAvailableNow, | ||
| // so no more listings need to happen | ||
| source.latestOffset(FileStreamSourceOffset(-1L), ReadLimit.maxFiles(5)) |
HeartSaVioR
changed the title
|
Thanks! Merging to master. |
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Outdated suggestions cannot be applied.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.
What changes were proposed in this pull request?
Files don't need to be cached for reuse in
FileStreamSourcewhen usingTrigger.AvailableNowbecause all files are already cached for the lifetime of the query inallFilesForTriggerAvailableNow.Why are the changes needed?
As reported in https://issues.apache.org/jira/browse/SPARK-44924 (with a PR to address #45362), the hard coded cap of 10k files being cached can cause problems when using a maxFilesPerTrigger > 10k. It causes every other batch to be 10k files, which can greatly limit the throughput of a new streaming trying to catch up.
Does this PR introduce any user-facing change?
Every other streaming batch won't be 10k files if using Trigger.AvailableNow and maxFilesPerTrigger greater than 10k.
How was this patch tested?
New UT
Was this patch authored or co-authored using generative AI tooling?
No