MinutesOct18

Introduction

• Attendees: Anne, Alain, Steve, Suchandra, Rob, Gabriele, John, Horst, Karthik, Arvind, Gabriele, Jeff, Xin, Rob, Tom, Ted, Terrence, Martin, John, Charles Bacon, Abhishek
• Apologies: none
• Coordinates: Thursdays, 2:30pm Central; 510-665-5437, #1212
• Previous meetings, MeetingMinutes
• ITB 0.7 background
  • ReleaseDocumentation
  • SiteValidationTable

Status of ITB 0.7.1 (Suchandra)

• Packages and cache - any changes needed?
  • follow-up potential bug in the configuration of the certs updater (Suchandra and Terrence are investigating)
    • Working now at UC_ITB, also for Terrence. Perhaps a transient error.
    • Will be putting in a change to configure-osg for glexec and configure squid. By Monday. Suchandra will test for these changes, and John W will.
  • Freezed date is now Oct 23
• VDT update to VDT 1.8.1b. Patch to Globus that fixes minor problems; several security fixes (Java, Tomcat, Voms Admin). Will do a VTB round before.
• Sites:
• FNAL_FERMIGRID_ITB
  • Follow-up: Encountering problem with gip and pbs (see VDT ticket 2959 ) Fixed (will be in the update) Y%
  • D0 succeeded on Fermigrid
• BNL_ITB_Test1
  • CMS failed w/ srmcp. Xin trying to get track this down (works on WN).
  • iptables fixed.
  • RSV permissions issue - fixed w/ Arvind.
  • Engage - had failed before - they will try again tomorrow.
  • Nanohub is failing as well. Xin thinks he fixed the problem (enabled).
• UC_ITB
  • Still working with Britta @ LIGO - Pegasus worflows not working.
  • GIP validator - default SE a requirement? Need to consult w/ Burt.
• CIT_ITB_1
  • Engage had failed, but probably have since succeeded.
• TTU-TESTWULF
  • No updates, no recent complaints, all appears okay.
• UCSDT2-ITB1
  • Waiting for Engage, D0; LIGO seems to be working.
  • Follow-up: Saw problems with cert updater Resolved - was an ordering problem in the documentation for privilege sites.

VO validation (Abhishek)

• Follow-up: Engage problems on a few sites (UCSD, CIT -- both responded, waiting on BNL), other sites were run successfully
• Starting tomorrow - will start putting in green flags on validated sites, and start wrapping up the validation.
• D0SAR? tried on OUHEP - their tests were successful.

Report from attributes meeting (Suchandra, Gabriele, Anand)

• See minutes at MinutesVTBOct16
• Need to get more experience in production use to better define the critical attributes.
• Went over ITB sites to check validity of attributes. Two sites have problems - FNAL_FERMIGRID_TEST (not up to date). For UCSDT2-ITB1 condor-status not being run since it wasn't a user.
• Need to investigate an issue with Class Ad evaluation in the GIP validator.
• GUMS service entities query, and translation by CEMon to BDII may have a problem. Is there a problem with the aggregator filtering non-registered ITB sites. Why does the BDII retaining old information.
• GIP configuration issue w/ CEMon
  • triage config issue to GIP support issues to Will
  • something that can be improved with standard trouble shooting notes to be included in the release documentation

WS-GRAM testing deployment

• Follow-up on these issues
• VOs willing to run workflows through WS-GRAM (Suchandra, Jeff)
  • STAR * Have been working with Star developers - looks straightforward. Having problems with production clients at Brookhaven. Older version of Condor 6.8.4. * Firewall problems with the OSG 0.6 client.
  • LIGO * Confirmed w/Britta the ability to run Pegasus through WS-Gram
• Replicating earlier ws-gram work against ITB sites (Jeff)
  • started this work last week.
  • Half the sites are working fine, the other are having small problems.
• Core container issues that Gram relies on - Charles w
• Does Gratia work correctly with WS-Gram if the condor 6.9? Need to address this.
• Documentation updated.

RSV status (Arvind)

• RSV SAM reporting status update
  • Sample record uploaded from OSG side to SAM.
  • Efforts ongoing between Arvind and James/David/Piotr to automate process from ITB RSV collector to other end. On our end, we've clarified what needs to be sent and will write code to make it happen; on their side they need to do some code mods to allow data that conforms to the WLCG specs 0.91 (that we'll be sending -- their system is not tweaked fully to comply with the specs yet).
  • Arvind will write a debugging guide for failures seen in the SAM console.
  • From CE point of view, no changes required in advance of OSG 0.8 deployment.
• Any follow-up on the issue:
  • Condor-devel brought in because production condor installations (6.8.x) don't support cron functionality; RSV should use existing condor if user points to the a condor 6.9.x installation; In general case condor should work, but there might be a conflict between ManagedFork and RSV since both use the local universe; Scot Kronenfeld thinks this might not work, but Terrence will try it out on a system to check
    • Terrence reported back to the ITB list that setting certain environment variables before pulling in the OSG-RSV package, as documented in the twiki, worked fine for him. RSV was using his existing 6.9.x condor install following that.
• Arvind going on vaction next Thursday for three weeks.
  • Will plan to have infrastructure in place for the SAM before
  • Scott Kronefeld will be the point of contact for RSV
• Xin: at BNL experiencing a problem restarting RSV, not removing stale Condor processes. Unique installation? Scott investigating.

OSG 0.8 Deployment readiness (RobQ)

• bdii - schema update required
• bdii-aggregator (no change) and monitor (no change)
  • http://is.grid.iu.edu/cgi-bin/status.cgi ?
• gip validator, http://gip-validate.grid.iu.edu/ (Consult w/ Will and Anand)
• rsv (ready)
• rsv-->sam (in progress)
• vors (no change) will continue for the foreseeable future
• ress? (No changes required)
• ress-validator? (will activate once deployment sites arrive)
• gratia? No changes required.
• voms-monitor, http://voms-monitor.grid.iu.edu/cgi-bin/index.cgi (no change required)
• preparation of top-level webs, http://www.opensciencegrid.org/Grid_Monitoring (remove gridcat)
• other services? (none)

OSG sites survey

• See links of results so far at SitesCoordinationPlanning

Release documentation (Rob, Anne)

• Doc-fest meeting this Thurs/Friday at Fermilab
• ReleaseDocumentation will become ReleaseDocumentation for OSG 0.8
• ReviewStatus replaces DocumentationTable
• Need volunteers for "responsibles" and "reviewers"

AOB

• dccp in workernode client? A request from CMS and ATLAS. Alain will consult Ted.

-- RobGardner - 17 Oct 2007