You are here: Home User Information Facility Services Frontier Frontier Meetings for ATLAS Meeting Minutes Minutes: 12/15/2011

Minutes: 12/15/2011

by John S. De Stefano Jr. last modified Dec 28, 2011 02:15 PM
Notes from the ATLAS Frontier meeting on December 15, 2011.

frontier-minutes-20111215.txt — Plain Text, 2 kB (2866 bytes)

File contents

Participants: Florentin Bujor, John DeStefano, Alastair Dewhurst, David Front, Dave Dykstra, Emmanouil Vamvakopoulos, Alexei Sedov
Apologies: Dario Barberis, Fred Luehring 

*** Site news and issues: ***

IN2P3:
- New Frontier node tested. Needs to be added to TiersofATLAS.py (This was done after meeting https://savannah.cern.ch/bugs/index.php?89940)
- E.V. has noticed that some NorduGrid sites are using 1 of the nodes at Lyon.  This was probably as a backup because of the network downtimes RAL have experienced.  We need to identify how to change this as NorduGrid are not in TiersofATLAS.py

BNL:
- Production is fine.
- Request has been placed to modify TiersofATLAS.py to use the new alias pointing at the new hardware. (This was done during the meeting!)

CERN:
- No news.  Everything seems to be working.

RAL:
- Network problems recently have caused sites to fail-over but no jobs appear to fail.  Otherwise working well.
- Started testing new RPMs on test box lcg0624.gridpp.rl.ac.uk.

*** Development: ***

RPMs:
- Something wrong in awstats configuration, small amount of data being lost.
- If you are doing RPM testing you should use:
  * frontier-squid-2.7.STABLE9-5.19.sl5 
  * frontier-tomcat-6.0.35_3.29-1 
  * frontier-awstats-6.0-14

AGIS integration:
- GGUS tickets submitted to sites with incorrect SITE_NAME variable
- OSG Site names have now been corrected by Xin in TiersofATLAS.py
- Work still needs to be done creating the scripts for modifying AGIS.
- Dave has asked how it is possible with AGIS to not use an alias such that the client can do the balancing.  eg., people from outside would use the alias where as WN at the site may use individual machines.

Monitoring:
- Large awstats files are still problematic.  A script has been produced to reduce the size of the log files.  However you also need to be running the latest RPMs to stop producing the large files.  There are 3 options:
1) Update to latest RPMs and run cleaning script once.
2) Deploy patch to old RPMs and run cleaning script once.
3) Run cleaning script regularly.
RAL - Alastair will update
BNL - New machines have latest RPMs (John will run cleaning script)
Lyon - Current machines have latest RPMS (Emmanouil will run cleaning script)
CERN - Florentin will investigate what needs to be done.  Unlikely he will be able to update before Christmas so may choose option 3.
- Request has been placed to remove pic and add T0 Frontiers from TiersofATLAS.py
- There is a new page on SSB for Frontier/squid monitoring (http://dashb-atlas-ssb.cern.ch/dashboard/request.py/siteview#currentView=Frontier_Squid&highlight=false)
- New BNL machines need to be added into awstats

*** A.O.B.: ***
Discussed meeting times.  Either Wednesday or Thursday is ok.  Will have to find out what meeting it is easier to alternate with.
First meeting of 2012 will be 12th/13th January.
Document Actions
Filed under: , , ,