[HADOOP-1106] Number of DataNodes less than target replication causes NameNode WARN message every millisecond

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Cannot Reproduce
• Affects Version/s: 0.12.0
• Fix Version/s: 0.13.0
• Assignee: Hairong Kuang
• Reporter: Nigel Daley
• Created: 09/Mar/07 23:09
• Updated: 08/Jul/09 16:42
• Resolved: 16/May/07 18:37

Description

When I startup a 4 DataNode cluster that already contains files with a target replication of 3, but 2 of the DataNodes fail to come online (ie. there are fewer DataNodes then the target replication), I get the following output in the NameNode log every millisecond or less, continued forever:

2007-03-09 22:45:22,848 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,849 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,850 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,851 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,852 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,852 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,853 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,853 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2
2007-03-09 22:45:22,854 WARN org.apache.hadoop.fs.FSNamesystem: Not able to place enough replicas, still in need of 2