HyukjinKwon commented on a change in pull request #23438: [SPARK-26525]Fast
release ShuffleBlockFetcherIterator on completion of the iteration
URL: https://github.com/apache/spark/pull/23438#discussion_r251227558
##########
File path:
core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala
##########
@@ -212,6 +214,9 @@ final class ShuffleBlockFetcherIterator(
logWarning("Failed to cleanup shuffle fetch temp file " + file.path())
}
}
+ // Unregister completion callbacks to avoid memory leak for
ShuffleBlockFetcherIterator
+ // contains lots of metadata about MapStatus
+ context.removeTaskCompletionListener(cleanupListener)
Review comment:
```
[error] * abstract method
removeTaskCompletionListener(org.apache.spark.util.TaskCompletionListener)org.apache.spark.TaskContext
in class org.apache.spark.TaskContext is present only in current version
[error] filter with:
ProblemFilters.exclude[ReversedMissingMethodProblem]("org.apache.spark.TaskContext.removeTaskCompletionListener")
```
If `TaskContext` is meant to be an API, old classes that implemented
`TaskContext` will fail here in newer Spark since that doesn't implement
`removeTaskCompletionListener`. Looks it's not meant to be an API tho.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
