Github user zsxwing commented on a diff in the pull request:
https://github.com/apache/spark/pull/1542#discussion_r15283935
--- Diff: core/src/main/scala/org/apache/spark/MapOutputTracker.scala ---
@@ -130,7 +130,7 @@ private[spark] abstract class MapOutputTracker(conf:
SparkConf) extends Logging
if (statuses == null) {
logInfo("Don't have map outputs for shuffle " + shuffleId + ",
fetching them")
var fetchedStatuses: Array[MapStatus] = null
- fetching.synchronized {
+ shuffleId.toString.intern.synchronized {
--- End diff --
This is not a correct fix for the following reasons:
* Should always use the same lock to protect the access of `fetching` to
set up the happen-before relation.
* `fetching.wait()` should always be used under the protection of the
`fetching` lock. See javadocs:
http://docs.oracle.com/javase/7/docs/api/java/lang/Object.html#wait()
---
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.
---
