[HDFS-10733] NameNode terminated after full GC thinking QJM is unresponsive.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.6.4
• Fix Version/s: 2.8.0, 2.9.0, 2.7.4, 3.0.0-alpha2
• Component/s: namenode, qjm
• Target Version/s: 2.7.4
• Hadoop Flags: Reviewed
• Assignee: Vinitha Reddy Gankidi
• Reporter: Konstantin Shvachko
• Created: 08/Aug/16 23:34
• Updated: 18/Aug/17 22:02
• Resolved: 18/Jan/17 21:30

Description

NameNode went into full GC while in AsyncLoggerSet.waitForWriteQuorum(). After completing GC it checks if the timeout for quorum is reached. If the GC was long enough the timeout can expire, and QuorumCall.waitFor() will throw TimeoutExcpetion. Finally FSEditLog.logSync() catches the exception and terminates NameNode.

Issue Links

• relates to: HDFS-12323 NameNode terminates after full GC thinking QJM unresponsive if full GC is much longer than timeout