Thank you Cristina
I've opened the follow GGUS
https://ggus.eu/ws/ticket_info.php?ticket=91260
My italian is not so go but from the developer answer (attached to
your email) he states that we should deploy/use an emi-2 WMS. EMI-1 is
still supported, right?! So, I hope that the answer to the GGUS from
the developpers is not "please deploy a EMI-2 WMS".
Thanks for the reply
Cheers
Goncalo
Quoting "Cristina Aiftimiei" <[log in to unmask]>:
> Hi Gonçalo,
>
> from the developers: DAG support was completely restructured in EMI 2
> version, where condor is no more used, because the previous design had
> some limitations and bugs. The EMI 2 version has to be used.
>
> Please use the GGUS tkt-ing system for more details and support from
> the developers side.
>
> Thank you,
> Cristina
>
>
> I DAG sono stati completamente ristrutturati in emi-2, dove non si
> usa più condor, perché il design precedente aveva diversi limiti, e
> bachi.
> Devono usare il wms di emi-2.
> Gonçalo Borges wrote:
>> Dear WMs experts ...
>>
>> A user is sending several dag jobs to my WMS (and LB) both in EMI 1.
>>
>> However, he is complaining that while all the children jobs are
>> RUNNING or DONE, the father is always reporting its status as
>> WAITING. In appendix I'm sending more details from a logging info
>> command.
>>
>> I've been through some FAQs like
>> https://wiki.egi.eu/wiki/Tools/Manuals/TS55
>> which do not seem to apply because it has passed enough time for
>> the WMS to recover from any long backlog or sporadic load.
>>
>> The only logs I found in the WMS are:
>>
>> # grep XqVdHXrO7aV5f3nH_0fzFw /var/log/wms/wmproxy*
>> /var/log/wms/wmproxy.log:04 Feb, 15:34:12 -I- PID: 22036 -
>> "wmpcoreoperations::regist DAG": Registering job id:
>> https://lb01.ncg.ingrid.pt:9000/XqVdHXrO7aV5f3nH_0fzFw
>>
>> # grep XqVdHXrO7aV5f3nH_0fzFw /var/log/wms/workload_manager_events.log
>> 04 Feb, 15:35:18 -I: [Info]
>> operator()(/builddir/build/BUILD/glite-wms-manager-3.3.1/build/../src/dispatcher_utils.cpp:228): new jobsubmit for
>> https://lb01.ncg.ingrid.pt:9000/XqVdHXrO7aV5f3nH_0fzFw
>> 04 Feb, 15:46:40 -I: [Info]
>> operator()(/builddir/build/BUILD/glite-wms-manager-3.3.1/build/../src/submit_request.cpp:338): https://lb01.ncg.ingrid.pt:9000/XqVdHXrO7aV5f3nH_0fzFw, collection
>> delivered
>>
>> I'm a bit confused because I'm not able to find the information
>> that allows me to follow up the job.
>>
>> Any help is appreciated in understanding why the WMS keeps saying
>> the job is WAITING while all the childdren are DONE.
>>
>> Thank you in advance
>> Cheers
>> Goncalo
>
>
> --
> Cristina Aiftimiei - EMI Project
> Ist. Naz. di Fisica Nucleare - Padova Address: via F. Marzolo, 8 -
> 35131 Padova - ITALY Phone: +39.049.967.7362
> Mobile: +39.3460230488
>
|