mridulm commented on a change in pull request #33446:
URL: https://github.com/apache/spark/pull/33446#discussion_r780737708
##########
File path:
core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala
##########
@@ -334,6 +353,7 @@ final class ShuffleBlockFetcherIterator(
s"due to Netty OOM, will retry")
}
remainingBlocks -= blockId
+ logFetchOnCompletionIfSlow()
Review comment:
If we are calling `logFetchOnCompletionIfSlow` from here, we should
recompute what is `totalRequestSize` and `blockCount` - since fetch failed
before all blocks were fetched.
Essentially, we need to compute the stats based on infoMap.keys --
remainingBlocks
Given the failure, `remainingBlocks.isEmpty` would never be empty except if
this was last block - potentially loosing out on all suffix requests taking
time.
--
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]
