r7raul1984 opened a new pull request, #3780:
URL: https://github.com/apache/celeborn/pull/3780

   ## Summary
   
   When a Celeborn shuffle fetch failure occurs, the resulting 
`FetchFailedException` message previously did not include any information about 
which worker the fetch was attempted from. This made it difficult to correlate 
failures with specific workers during incident investigation.
   
   This PR propagates the worker hostname into the `FetchFailedException` 
message so that it appears in Spark's task failure logs and event history.
   
   ## Changes
   
   ### `ExceptionMaker.java`
   Added `String workerHost` parameter to the `makeFetchFailureException` 
interface method:
   ```java
   // before
   Exception makeFetchFailureException(int appShuffleId, int shuffleId, int 
partitionId, Exception e);
   
   // after
   Exception makeFetchFailureException(int appShuffleId, int shuffleId, int 
partitionId, String workerHost, Exception e);
   ```
   
   ### `CelebornInputStream.java`
   Resolved the worker hostname from `currentReader.getLocation().getHost()` 
(falling back to `"unknown"`) and passes it to `makeFetchFailureException`.
   
   ### `ExceptionMakerHelper.scala` 
   Updated both implementations to accept and embed `workerHost` in the 
exception message:
   ```
   Celeborn FetchFailure with appShuffleId/shuffleId: <id>/<id> from worker 
<host>
   ```


-- 
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]

Reply via email to