[HDFS-3264] race in removing error edit stream could cause fatalExit()

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.1.0
• Component/s: namenode
• Assignee: Brandon Li
• Reporter: Brandon Li
• Created: 12/Apr/12 16:23
• Updated: 12/Apr/12 16:23

Description

An unhealthy storage directory and its edit stream can be removed from editStreams list. There are multiple places to remove unhealthy directory and its stream, such as log sync, checkpointing.

Method removeEditsStreamsAndStorageDirs() could unneceissarily invoke fatalExit() if another thread removes one error editStream before current thread. This race could be between logSync requests, or between logSync and any place where a storage directory may be removed, e.g., during checkpointing.