Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
1.2.1
-
None
Description
DiskBlockManager registers a Runtime shutdown hook, which creates a hard reference to the entire Driver ActorSystem. If a long-running JVM repeatedly creates and destroys SparkContext instances, it leaks memory.
I suggest we deregister the shutdown hook if DiskBlockManager.stop is called. It's redundant at that point.
PR coming.
Attachments
Issue Links
- is duplicated by
-
SPARK-5869 Exception when deleting Spark local dirs when shutting down DiskBlockManager
- Resolved
-
SPARK-8024 Luigi triggering resolved Blockmanager bug
- Resolved
- links to