And there must be another exception earlier in the log. Since this NPE happens at the job update time only shows missing mapred counter, there shall be a real error that causes the mapred failure ahead of this exception.
On Sat, Feb 7, 2015 at 11:20 PM, 周千昊 <[email protected]> wrote: > Hi, dong > Can you please provide the exact commit of which you are using? > > On Thu Feb 05 2015 at 2:45:34 PM dong wang <[email protected]> wrote: > > > as checked in the apache-tomcat-7.0.57/logs/kylin_job.log, it says as > > below: > > > > [QuartzScheduler_Worker-9]:[2015-02-05 > > 14:15:22,572][DEBUG][com.kylinolap.job.tools.HadoopStatusChecker. > > checkStatus(HadoopStatusChecker.java:74)] > > - State of Hadoop job: job_1422357952310_0226:FINISHED-FAILED > > [QuartzScheduler_Worker-9]:[2015-02-05 > > 14:15:22,622][ERROR][com.kylinolap.job.cmd.JavaHadoopCmdOutput. > > updateJobCounter(JavaHadoopCmdOutput.java:176)] > > - > > java.lang.NullPointerException > > at > > com.kylinolap.job.cmd.JavaHadoopCmdOutput.updateJobCounter(J > > avaHadoopCmdOutput.java:162) > > at > > com.kylinolap.job.cmd.JavaHadoopCmdOutput.getStatus(JavaHado > > opCmdOutput.java:85) > > at > > com.kylinolap.job.flow.AsyncJobFlowNode.execute(AsyncJobFlowNode.java:86) > > at org.quartz.core.JobRunShell.run(JobRunShell.java:202) > > at > > org.quartz.simpl.SimpleThreadPool$WorkerThread.run( > > SimpleThreadPool.java:573) > > >
