Re: Lot of ClosedChannelExceptions in namenode

2009-10-02 Thread Murali Krishna. P
org.apache.hadoop.hdfs.server.namenode.FSNamesystem: blk_7800017541867395810_15891336 is beening recovered, igno ring this request. Thanks, Murali Krishna From: Raghu Angadi To: common-user@hadoop.apache.org Sent: Fri, 2 October, 2009 11:07:16 AM Subject: Re: Lot of ClosedChannelExceptions in

Re: Lot of ClosedChannelExceptions in namenode

2009-10-01 Thread Raghu Angadi
This is more likely a symptom than cause. One possible reason is that clients are dying or restarting. Looks like "caught: ..." log messages include stack trace as well, could you include couple of those? number of handlers might not change the situation. Even if the number of handlers is not eno