zhan7236 opened a new pull request, #2688:
URL: https://github.com/apache/uniffle/pull/2688
### What changes were proposed in this pull request?
This PR disables the `dataPusher` initialization for Spark driver, as it's
only needed for executors to push shuffle data.
### Why are the changes needed?
The `dataPusher` is used to push shuffle data to shuffle servers, which is
an executor-side operation. Driver does not push shuffle data, so initializing
`dataPusher` (along with its thread pool) for driver is unnecessary and wastes
resources.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
- Compiled successfully for both Spark 2 and Spark 3 profiles
- All existing tests pass:
- `RssShuffleManagerTest` (Spark 3): 9 tests passed
- `DelegationRssShuffleManagerTest` (Spark 3): 8 tests passed
- `DelegationRssShuffleManagerTest` (Spark 2): 5 tests passed
### Summary of Changes:
1. **RssShuffleManagerBase.java**: Wrapped `dataPusher` initialization (both
`DataPusher` and `OverlappingCompressionDataPusher`) with `if (!isDriver)`
condition
2. **RssShuffleManager.java (Spark 3)**: Added null check for
`dataPusher.setRssAppId()` calls in `registerShuffle()` and `setPusherAppId()`
methods
3. **RssShuffleManager.java (Spark 2)**: Added null check for
`dataPusher.setRssAppId()` calls in `registerShuffle()` and `getWriter()`
methods
Closes #1603
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
