[HDFS-5346] Avoid unnecessary call to getNumLiveDataNodes() for each block during IBR processing

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.23.9, 2.3.0
• Fix Version/s: 0.23.10, 2.3.0
• Component/s: namenode, performance
• Hadoop Flags: Reviewed
• Assignee: Ravi Prakash
• Reporter: Kihwal Lee
• Created: 11/Oct/13 14:36
• Updated: 03/Sep/14 23:41
• Resolved: 16/Oct/13 21:20

Description

When initial block reports are being processed, checkMode() is called from incrementSafeBlockCount(). This causes the replication queues to be initialized in the middle of processing a block report in the IBR processing mode. If there are many block reports waiting to be processed, SafeModeMonitor won't be able to make name node leave the safe mode soon. It appears that the block report processing speed degrades considerably during this time.

Update: The main issue can be resolved by config. The other issue of calling getNumLiveDataNodes() for each block in the block report will be addressed in this jira