Hi Marcel, the ce is in the bdii, but just to be sure before checking
more complex things, which VO do you use?
the only ones supported by that queue are: dech,dgtest,ghep,hp
Cheers,
Daniele
Marcel Schroers wrote:
> Hi Daniele, Hi Maarten,
>
> thanks for you help.
>
> I've just restarted the service, by now the CE
> is not appearing, maybe it needs some time.
>
> Our II_Contact is:
> II_Contact = "bdii-fzk.gridka.de";
>
> If we are not listed in the BDII then what could leed to such a
> behavior? Should I recheck our configuration of the CE/WMS(but i
> didn't reconfigured the machines in the past two weeks and last
> week everything seemed fine)? And what should I look for then?
>
> cheers,
> marcel
>
>
> Am 11.04.2008 um 12:41 schrieb Daniele Cesini:
>> Hi Marcel,
>>
>> Marcel Schroers wrote:
>>> Hi Daniele,
>>>
>>> first of all: thanks for the reply.
>>>
>>> - I've raised the timeout value in the glite_wms.conf file, it was
>>> indeed set to the
>>> default value of 30s. (do I have to restart any services?)
>> yes, /opt/glite/etc/init.d/glite-wms-wm restart
>>>
>>> - then I've removed my Requirements line (it points to our
>>> grid-ce.physik.uni-wuppertal.de)
>>> and submitted the job. And voila the job ran at sara.nl.
>>>
>>> - a glite-job-list-match with the Req. Line in the jdl file returns:
>>> ---
>>> ~/panda-get$ > glite-job-list-match filetrans.jdl
>>>
>>> Selected Virtual Organisation name (from proxy certificate
>>> extension): atlas
>>> Connecting to host glite-wms.physik.uni-wuppertal.de, port 7772
>>>
>>>
>>> ===================== glite-job-list-match failure
>>> ======================
>>> No Computing Element matching your job requirements has been found!
>>> ======================================================================
>>> ---
>>> So there seems to be a problem between the CE and the WMS, but still
>>> I'm a bit clueless.
>> The problem is probably that the required ce is not present in the
>> BDII used by the wms. You can find the bdii in the glite_wms.conf
>> file, II_Contact parameter. Could you paste the bdii host?
>>
>>> Does raising the timeout maybe fix the problem? Or can it also be
>>> another problem?
>>>
>> no, if the ce is not in the bdii the increased timeout won't help.
>>
>> Cheers,
>> Daniele.
>>> Thanks in advance for all help.
>>>
>>> cheers,
>>> marcel
>>>
>>>
>>> Am 11.04.2008 um 12:01 schrieb Daniele Cesini:
>>>> Hi, I think that the problem is that zero CE are matching the job
>>>> requirements.
>>>> Did you try a listmatch first?
>>>> This can be due to wrong (or too strict) requirements in jdl or to
>>>> empty information supermarket (the snapshot of the grid resource
>>>> available for job submission) on the WMS.
>>>> The information supermarket sometimes gets empty because of the too
>>>> low value for the II_timeout parameter in the glite_wms.conf file.
>>>> It should be greater than the default 30s (90s is a good value).
>>>> Another possibility is that the bdii pointed by the wms is not
>>>> publishing correctly.
>>>> Do you have access to the WMS?
>>>> Cheers,
>>>> Daniele.
>>>>
>>>> Marcel Schroers wrote:
>>>>> Hello everybody,
>>>>>
>>>>> we at wuppertal are currently having some problems with
>>>>> our WMS Server. Submitting a job via glite-job-submit
>>>>> results in the following error message:
>>>>>
>>>>> ---
>>>>> ~$ > glite-job-status -i jobid
>>>>>
>>>>>
>>>>> *************************************************************
>>>>> BOOKKEEPING INFORMATION:
>>>>>
>>>>> Status info for the Job :
>>>>> https://glite-wms.physik.uni-wuppertal.de:9000/zJG2uck3p0xZm7cSUYmsJQ
>>>>> Current Status: Waiting
>>>>> Status Reason: BrokerHelper: Problems during rank evaluation
>>>>> (e.g. GRISes down, wrong JDL rank expression, etc.)
>>>>> Submitted: Fri Apr 11 09:29:43 2008 CEST
>>>>> *************************************************************
>>>>> ---
>>>>>
>>>>> A submission via the lcg-RB works fine. I've searched in the
>>>>> archives of the LCG-ROLLOUT Mailinglist,
>>>>> but all suggestions there didn't solve the problem. I spoke to
>>>>> Torsten Harenberg (who is currently
>>>>> at CERN) and he pointed out that it maybe could be a database
>>>>> problem. But neither him nor me have
>>>>> an idea where the database could be or if this will solve the
>>>>> problem.
>>>>> Any suggestions are welcome.
>>>>>
>>>>> Thanks in advace for all help.
>>>>>
>>>>>
>>>>> cheers,
>>>>> marcel
>>>>>
>>>>>
>>>
>>>
>>> ((((((((((((((((((It's not stupid, it's advanced))))))))))))))))))
>>> ( )
>>> ( Marcel Schroers [log in to unmask] )
>>> ( Bergische Universitaet )
>>> ( FB C - Physik Tel.: +49 (0)202 439-2631 )
>>> ( Gaussstr. 20 )
>>> ( 42097 Wuppertal )
>>> ((((((((((((((((((((((((((((((((()))))))))))))))))))))))))))))))))
>>
|