[
https://issues.apache.org/jira/browse/SPARK-2896?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14088818#comment-14088818
]
Apache Spark commented on SPARK-2896:
-------------------------------------
User 'sarutak' has created a pull request for this issue:
https://github.com/apache/spark/pull/1823
> Remove redundant error notification by results.put() when fetching from remote
> ------------------------------------------------------------------------------
>
> Key: SPARK-2896
> URL: https://issues.apache.org/jira/browse/SPARK-2896
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 1.0.1
> Reporter: Kousuke Saruta
> Priority: Minor
>
> In BasicBlockFetcherIterator#sendRequest, there is a part of code as follows.
> {code}
> case Failure(exception) => {
> logError("Could not get block(s) from " + cmId, exception)
> for ((blockId, size) <- req.blocks) {
> results.put(new FetchResult(blockId, -1, null))
> }
> }
> {code}
> This is a part of code for notifying error has occurred.
> In this code, results.put is called for all of req.blocks but I think it's
> redundant; only first error is enough.
> It is because when .iterator.next() returns error result, then it throws
> FetchFailedException immediately.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
