Ali, i saw that we already have a bug for that. See http://bugzilla.globus.org/globus/show_bug.cgi?id=4908 Because i cannot reproduce it: Could you please check if the fix of Dominic works for you too and give us feedback once you're done? Thanks, Martin
> > Exactly. This seems to be the cause of the problem.... and the OS which we > use is : 3.0.8 version of Scientific Linux- CERN Release. > /Ali > > >> The problem seems to be >> "Can't fetch initial working directory at >> /usr/local/globus-4.0.5/lib/perl/Globus/GRAM/JobManager.pm" >> in cache cleanup in non-detached mode. I didn't see that before >> and currently can't say what the problem is. >> Need to look more into it. >> What operating system are you using? >> >> Martin >> >>> Hi Charles, >>> I've attached both container logs with this job submission command line >>> : >>> "globusrun-ws -submit -streaming -F >>> https://130.237.221.105:8444/wsrf/services/ManagedJobFactoryService -c >>> /bin/date" >>> ----------------------------------------------------------------------- >>> For the detached container I got back this results : >>> ----------------------------------------------------------------------- >>> Delegating user credentials...Done. >>> Submitting job...Done. >>> Job ID: uuid:da5ae1fe-8de2-11dc-ab8d-00188b25ea22 >>> Termination time: 11/09/2007 10:10 GMT >>> Current job state: Active >>> Current job state: CleanUp-Hold >>> Thu Nov 8 11:10:43 CET 2007 >>> Current job state: CleanUp >>> Current job state: Done >>> Destroying job...Done. >>> Cleaning up any delegated credentials...Done. >>> ----------------------------------------------------------------------- >>> And for non-detached the following: >>> ----------------------------------------------------------------------- >>> Delegating user credentials...Done. >>> Submitting job...Done. >>> Job ID: uuid:84b8bc3e-8de3-11dc-84a5-00188b25ea22 >>> Termination time: 11/09/2007 10:15 GMT >>> Current job state: Active >>> Current job state: CleanUp-Hold >>> Thu Nov 8 11:15:26 CET 2007 >>> Current job state: CleanUp >>> Current job state: Failed >>> Destroying job...Done. >>> Cleaning up any delegated credentials...Done. >>> ----------------------------------------------------------------------- >>> >>> Best regards and thanks >>> Ali >>> >>> >>> >>> >>> >>>> For debugging, I would request: >>>> >>>> 1) globusrun-ws commands *without* -debug. It just adds SOAP >>>> messages, which usually don't help debug. >>>> >>>> 2) Container logs for both kinds of job, one that succeeds, and one >>>> that fails. Your gram-debug.log only had one of the two job uuids >>>> whose clientside traces you provided. >>>> >>>> So, to be clear, the logs would be: >>>> 1) jobA to detached container >>>> 2) the corresponding container log >>>> 3) jobB to the non-detached container >>>> 4) the corresponding non-detached container log >>>> >>>> >>>> Charles >>>> >>>> On Nov 1, 2007, at 4:31 PM, Ali Gholami wrote: >>>> >>>>> Yes Martin, you are totally right.I just forgot to start the postgres >>>>> service and consequently it failed the job.I have just attached the >>>>> log >>>>> files of the both modes with the running postgres.The command >>>>> lines are >>>>> the same as before.You see that in detached mode , the result has >>>>> been >>>>> sent back. >>>>> >>>>> Thanks for your consideration >>>>> Ali >>>>> >>>>> >>>>> >>>>>> Ali, >>>>>> It seems like you didn't configure RFT properly. >>>>>> >>>>>> 2007-10-31 21:09:18,198 ERROR service.ReliableFileTransferImpl >>>>>> [main,<init>:69] Unable to setup database driver with >>>>>> pooling.Connection >>>>>> refused. Check that the hostname and port are correct and that the >>>>>> postmaster is accepting TCP/IP connections. >>>>>> >>>>>> Please check the quickstart quide at >>>>>> http://www.globus.org/toolkit/docs/4.0/admin/docbook/ >>>>>> quickstart.html#q-rft-configure >>>>>> for this. >>>>>> >>>>>> There seems to be another issue, but please fix the above first. >>>>>> Once you fixed that: >>>>>> Do you still see the difference regarding globus-start-container and >>>>>> globus-start-container-detached for >>>>>> * jobs without streaming and without file staging? >>>>>> * jobs without streaming and with file staging? >>>>>> * jobs with streaming like you did it >>>>>> >>>>>> (See >>>>>> http://www.globus.org/toolkit/docs/4.0/execution/wsgram/user- >>>>>> index.html >>>>>> for how to specify staging in the job description) >>>>>> >>>>>> Martin >>>>>> >>>>>> >>>>>>> Thank you very much for your answer.I have attached the log files >>>>>>> to >>>>>>> this >>>>>>> email.The commands that I used are as follwoing: >>>>>>> At globus side: " globus-start-container -p 8444 1>> gram- >>>>>>> debug.log 2>> >>>>>>> gram-debug.log " >>>>>>> >>>>>>> And for the job submission:"globusrun-ws -submit -streaming -dbg >>>>>>> -F >>>>>>> https://130.237.221.105:8444/wsrf/services/ >>>>>>> ManagedJobFactoryService -c >>>>>>> /bin/date 1>>globususer.log 2>>globususer.log" >>>>>>> >>>>>>> /Ali >>>>>>> >>>>>>> >>>>>>> >>>>>>>> Ali, >>>>>>>> >>>>>>>> We would need some more information about that: >>>>>>>> Please enable debug logging in the container (set >>>>>>>> log4j.category.org.globus=DEBUG in >>>>>>>> $GLOBUS_LOCATION/container-log4j.properties) >>>>>>>> >>>>>>>> You can store the output in non-detached mode e.g. with >>>>>>>> globus-start-container 1> file 2> file >>>>>>>> >>>>>>>> Also: Please submit the job using the debug option in >>>>>>>> globusrun-ws (-dbg) and store the output on the client side >>>>>>>> >>>>>>>> Please send output of both >>>>>>>> >>>>>>>> Martin >>>>>>>> >>>>>>>> >>>>>>>>> Hi folks, >>>>>>>>> Does any body knows the difference between "globus-start- >>>>>>>>> container" >>>>>>>>> and >>>>>>> "globus-start-container-detached" in GT-4.0.5. When I try to submit >>>>>>> jobs >>>>>>>>> in the first case, jobs are failed, but in the second mode , I >>>>>>>>> get >>>>>>>>> back >>>>>>> the results! >>>>>>>>> Thanks in advance for the answer >>>>>>>>> Ali >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> <globususer.log> >>>>>> <gram-debug.log> >>>>>> <detached-mode.log> >>>> >>>> >>> >> >> >> > > >
