Github user massie commented on a diff in the pull request:
https://github.com/apache/spark/pull/6423#discussion_r31542306
--- Diff:
core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala
---
@@ -290,22 +287,15 @@ final class ShuffleBlockFetcherIterator(
sendRequest(fetchRequests.dequeue())
}
- val iteratorTry: Try[Iterator[Any]] = result match {
+ val iteratorTry: Try[InputStream] = result match {
case FailureFetchResult(_, e) =>
Failure(e)
case SuccessFetchResult(blockId, _, buf) =>
// There is a chance that createInputStream can fail (e.g.
fetching a local file that does
// not exist, SPARK-4085). In that case, we should propagate the
right exception so
// the scheduler gets a FetchFailedException.
- Try(buf.createInputStream()).map { is0 =>
- val is = blockManager.wrapForCompression(blockId, is0)
- val iter =
serializerInstance.deserializeStream(is).asKeyValueIterator
- CompletionIterator[Any, Iterator[Any]](iter, {
- // Once the iterator is exhausted, release the buffer and set
currentResult to null
- // so we don't release it again in cleanup.
- currentResult = null
- buf.release()
- })
+ Try(buf.createInputStream()).map { inputStream =>
--- End diff --
The `ShuffleBlockFetcherIterator` has no information about server statuses
from the map output tracker, shuffle IDs, etc. Using `Try` allows the
`BlockStoreShuffleFetcher` to reformat exceptions as a `FetchFailedException`
which is the right exception to return to the scheduler.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
