[SPARK-3495] Block replication fails continuously when the replication target node is dead

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.0.2, 1.1.0
• Fix Version/s: 1.1.1, 1.2.0
• Component/s: Block Manager, DStreams, Spark Core
• Target Version/s: 1.1.1, 1.2.0
• Assignee: Tathagata Das
• Reporter: Tathagata Das
• Created: 12/Sep/14 01:01
• Updated: 11/Nov/14 02:37
• Resolved: 02/Oct/14 20:50

Description

If a block manager (say, A) wants to replicate a block and the node chosen for replication (say, B) is dead, then the attempt to send the block to B fails. However, this continues to fail indefinitely. Even if the driver learns about the demise of the B, A continues to try replicating to B and failing miserably.

The reason behind this bug is that A initially fetches a list of peers from the driver (when B was active), but never updates it after B is dead. This affects Spark Streaming as its receiver uses block replication.

Issue Links

• is duplicated by: SPARK-3498 Block always replicated to the same node
• links to: [Github] Pull Request #2366 (tdas)
• links to: [Github] Pull Request #3191 (tdas)