You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by "Murali Krishna. P" <mu...@yahoo.com> on 2010/10/14 15:29:58 UTC
Child Jvm issue
Hi,
After running few jobs sequentially, I get the following error in the
tasktracker log and the job fails. The new jvm is started and it exists
immediately. Because of this, no userlogs gets created for that particalar task.
The job also eventually fails. When I restart the workflow again, it runs fine
for sometime and gets into the same issue, (usually happend after the 3rd job is
started). All the jobs are similar with different inputs. What could be wrong? I
have tried increasing child java heap upto 2GB, but didnt work.
2010-10-14 06:17:36,576 INFO org.apache.hadoop.mapred.JvmManager: In JvmRunner
constructed JVM ID: jvm_201010140533_0015_r_-469452996
2010-10-14 06:17:36,576 INFO org.apache.hadoop.mapred.JvmManager: JVM Runner
jvm_201010140533_0015_r_-469452996 spawned.
2010-10-14 06:17:36,607 INFO org.apache.hadoop.mapred.JvmManager: JVM :
jvm_201010140533_0015_r_-469452996 exited. Number of tasks it ran: 0
The jobclient also gets errors like below which is mostly due the missing user
log because of the above issue.
Oct 14, 2010 6:17:55 AM org.apache.hadoop.mapred.JobClient getTaskLogs
WARNING: Error reading task
outputhttp://xxx.com:50060/tasklog?plaintext=true&taskid=attempt_201010140533_0015_m_000005_1&filter=stdout
Thanks for the help.
Thanks,
Murali Krishna