Hi, Thaks for the reply. I have solved the problem. I Got This message in the Globus Container Log, after I changed container-log4j.properties to enable GRAM debugging. Globus container continually display this message :
2008-06-14 09:01:00,055 DEBUG ManagedExecutableJobResource.9c51eec0-0f92-11dd-800a-edc3adbc375d [Thread-4,r emove:285] Waiting to be Done or Failed. Current state: None 2008-06-14 09:01:02,058 DEBUG ManagedExecutableJobResource.9c51eec0-0f92-11dd-800a-edc3adbc375d [Thread-4,g etInternalState:1666] getting resource datum internalState Globus seems always check the previous job that doesn't exist anymore. So I Check the folder .globus/persisted/192.168.0.3-8443 where the job id of that job stored. I deleted that folder and restart Globus container, and everything back to normal again. > Sorry, I'm not sure I understand what you mean by "The Job status > didn't changed". > > As far as I understand, Condor-G should eventually receive > notification that the jobs are done, and update their status. When > you checked with condor_q on the compute node, did you see that the > jobs were done? yes, i saw that the job has successfully executed by Condor. But At that time Globus still being Stopped. When Globus had started, It didn't send update about the jobs that already finished to Condor-G when Condor-G send request to update job status. > > You might also want to try the condor users list, as they are the > developers of condor-G and might know more about how it is meant to > behave. I know that condor-G creates a "gridmonitor" that keeps logs > about what it knows about grid jobs. I believe that is stored in a > subdirectory under /tmp, so you might want to check those logs to see > what is happening. > > > Charles > > On Jun 13, 2008, at 8:42 PM, [EMAIL PROTECTED] wrote: > >> >> Hi, >> I have a question about condor-G - GT Communication. >> Here is my scenario : >> >> I have 5 jobs that i submitted to Globus via Condor-G. When those jobs >> execute at Globus resource (Condor), I shutdown the Globus >> Container. The >> status of the jobs were : >> >> 473.0 tonny PENDING Condor blade3.bagong.grid >> /home/grid/mahasis >> 474.0 tonny PENDING Condor blade3.bagong.grid >> /home/grid/mahasis >> 475.0 tonny PENDING Condor blade3.bagong.grid >> /home/grid/mahasis >> 476.0 tonny ACTIVE Condor blade3.bagong.grid /home/grid/ >> mahasis >> 477.0 tonny PENDING Condor blade3.bagong.grid >> /home/grid/mahasis >> >> and 60 seconds after that, Those job were succesfully executed. But I >> didn't started globus immediately. I started globus container after >> 120 >> seconds. >> >> At this condition what will Condor-G do ? >> >> In my case The Job status didn't changed, although Condor-G probes the >> globus to update those jobs status. Does Anybody know why this >> condition can happen ? Thanks >> >> Tonny >> >> >> > >
