FYI, some of the analysis stress test over the grid.
SLAC is significant faster in "mean prepare inputs time" because (I think)
we are doing direct root file reading. It doesn't means our network IO is
faster, and these type of stress test didn't put high enough stress on our
storage.
Wei Yang | [log in to unmask] | 650-926-3338(O)
------ Forwarded Message
From: Nurcan Ozturk <[log in to unmask]>
Date: Thu, 23 Apr 2009 14:25:36 -0500 (CDT)
To: <[log in to unmask]>
Subject: [Usatlas-prodsys-l] first HammerCloud jobs for US cloud
Hello Tier1 and Tier2's,
The results of the first HammerCloud test with metrics were reported this
morning in the ADC Operations meeting, please see the slides at:
http://indico.cern.ch/getFile.py/access?subContId=1&contribId=3&resId=0&mate
rialId=slides&confId=57312
This report was on a test monitored at:
http://gangarobot.cern.ch/st/test_253/
Please check your site's status from the slides and from the monitoring
link. ANALY_SWT2_CPB was missed in the test, it'll be put in.
As you will see from the monitoring page, there is a list of input
datasets being used in the tests:
Input DS Patterns:
mc08.*Wmunu*.recon.AOD.e*_s*_r5*tid*
mc08.*Zprime_mumu*.recon.AOD.e*_s*_r5*tid*
mc08.*Zmumu*.recon.AOD.e*_s*_r5*tid*
mc08.*T1_McAtNlo*.recon.AOD.e*_s*_r5*tid*
mc08.*H*zz4l*.recon.AOD.e*_s*_r5*tid*
mc08.*.recon.AOD.e*_s*_r5*tid*
Please make it sure that you have a good replica of the matching datasets
at your site. Otherwise jobs will keep failing on the problematic files.
We will need to discuss in the next FacilityWGAP meeting on Tuesday how
often and at what scale we like to have HammerCloud to run in the US cloud.
Please also note on the slides that two mailing lists are setup now:
[log in to unmask]
[log in to unmask]
Please subscribe as appropriate.
Regards,
Nurcan.
_______________________________________________
Usatlas-prodsys-l mailing list
[log in to unmask]
https://lists.bnl.gov/mailman/listinfo/usatlas-prodsys-l
------ End of Forwarded Message
