Hi,all, I got this problem below every time I ran reduce. 10.254.106.48:50010:DataXceiver: java.io.IOException: Block blk_-7951711472001460544 is valid, and cannot be written to.
I am running map-reduce on *4* slave nodes,using *hadoop-0.16.4-core.jar* I was wondering whether this error is related with this issue(its *Affects Version/s:*0.14.1) https://issues.apache.org/jira/browse/HADOOP-1845 I changed "*mapred.submit.replication*" to *2* but still getting this error. Then I tried to use *3* nodes only,this time I still got all these error: 13 ERROR dfs.DataNode - 10.254.183.52:50010:DataXceiver: java.io.IOException: Block blk_-1748552216850253853 is not valid. /mnt/nutch/logs/hadoop.log:2008-08-09 23:29:47,794 WARN mapred.TaskTracker - Error running child /mnt/nutch/logs/hadoop-root-datanode-domU-12-31-39-00-88-02.log.2008-08-09:2008-08-09 21:59:02,025 ERROR dfs.DataNode - 10.254.143.240:50010:DataXceiver: java.io.IOException: Block blk_-2061672148590840392 is valid, and cannot be written to. /mnt/nutch/logs/hadoop-root-tasktracker-domU-12-31-39-00-88-02.log.2008-08-09:2008-08-09 23:29:47,807 WARN mapred.TaskTracker - Error from unknown child task: task_200808092153_0001_m_000006_0. Ignored. Any idea would be much appricaited. Thanks, Xu
