UKI Monthly Operations Meeting (TB-SUPPORT)

Thursday 23 Aug 2007, 10:30 → 11:30 Europe/Zurich

VRVS - Boat virtual room

Monthly review and discussion meeting for those involved with UKI EGEE/WCG deployment and operations.

Minutes - UKI Monthly Operations Meeting 23 September 2007 Site Status review and issues ============================= Accounting Irish sites have problem EFDA-JET - For some sites, problems with accounting records with multiple CEs and scaling factors See http://www.gridpp.ac.uk/wiki/RALPP_Logbook_200607 for how to do the scaling in torque Monitoring status ================= LHCb SAM problem - RHUL, Lancs, Shef RHUL has RedHat 7.3??!!?? Lancs looking at problem - but errors vague Availability Data ================= Strange things happening in the latter half of August with processing of tests Sites seem generally stable SL Atlas tests are broken at the moment Check FCR page Current experiment/VO Activites and issues ========================================== Atlas Now possible to move data off T2 sites, amount of work should increase CMS LHCb Regional VOs Gridpp vo very active Bio and pheno users causing problems, sites have been limited queed jobs <1000 Lancs reported Biomed problem 150 jobs 128-256Mb/s for 12h causing bandwidth problems WLCG Meetings ============= MB looking at pledges DTeam - What documentation need for Cluster security? MOU, Site response time, Need clear guidelines on how to decommision SEs. Killing low efficency jobs EGEE News ========= Concerns about DAG repository, JC to raise as issue with GDB gLite CE being dropped? Had not met criteria to remain, but has not been removed finally Handling Tickets ================ Tier 2 coordinators more involved in chasing tickets. AOB === Thanks to Philippa for her contribution MPI course in Dublin Sept 10-11 Next UKI meeting - 27 Sept

10:30 → 10:50 Site status review and issues

- Accounting -- csTCDie has problems since June -- cpDIASie has problems July -- Gap in EFDA-JET (site was down?) -- QMUL has stopped -- Cambridge has stopped (known APEL problem) - Monitoring status -- https://lcg-sam.cern.ch:8443/sam/sam.py -- Generally looks good under ops -- under LHCb the following have errors: RHUL; Lancaster; Sheffield - Site concerns - Availability data for August so far: http://www.gridpp.ac.uk/wiki/SAM_availability:_May-August_2007 - Steve Lloyd's ATLAS tests are broken at the moment! - Check FCR for your site when you have time: https://cic.in2p3.fr/index.php?section=rc&page=freedomofchoice

10:50 → 10:55 Current experiment/VO activities and issues

- ATLAS -- Data transfers are starting up again - CMS -- Preparing for CSA07. Sites involved will already be working with CMS. An rfio issue is stopping them using sites with SL4 WNs and DPM. - LHCb -- Simultation has just finished. Moving on to reconstruction (T1 task). Have problems with SL4 sites due to an lcg-cp error. This should be fixed this week. Had problems with software installations over the last month. Check out the LHCb SAM test pages. Good to check your site is not blacklisted. - Regional VOs -- gridpp is now quite active with many groups using it. Support it if you can. Regional VOs are not currently a top priority but there are still plans to set them up. - Other such as biomed -- Lots of problems over the last month from large temp files to job storms. Generally user gets banned and informed of problems being caused.

10:55 → 11:05 WLCG: Recent GDB/MB/DTEAM discussions

- There has not been a GDB since the last UKI meeting. The next GDB is a week Friday in conjunction with CHEP07: http://indico.cern.ch/conferenceDisplay.py?confId=17747. Is there anything you want raised at this meeting? - The MB is starting to look at Tier-2 performance against pledges. Data is being taken from APEL for this purpose. The UK looks low at 43% of pledge met but in fact the deployed CPU resources (8000+ KSI2K) are above the pledge and the issue is utilisation. We need sites to be more engaged in improving usage (where practical). - Recent DTEAM meetings have covered: what security contact information do we need to collect - what's missing in the documentation; the need to revisit the fairshares summary for all sites (http://www.gridpp.ac.uk/wiki/Current_VO_Fairshares_at_T2/T1); site response times - how do we measure it and what is possible; the need for clear guidelines on removing SEs; site availability; killing jobs and what the batch systems can report to the users.

11:05 → 11:15 EGEE news

- Suggestions for use cases for glue 2.0 (send ideas to Alessandra) - Voting for top priorities to be tackled by TCG - Use of the DAG repository for external dependencies (comments from ROCs are here: http://egee-docs.web.cern.ch/egee-docs/operational_tools/Operations_Meetings/2007/Responses_to_proposed_use_of_DAGS_repository_(up_to_Thurs_16_Aug).txt). - Possible decision to drop the gLite CE - SL4 progress for services: https://twiki.cern.ch/twiki/bin/view/EGEE/Glite31NodeTracker - GOCDB3 allows several settings for monitoring. This came up as there was confusion as to why some sites in downtime were still being monitored. -- Production + monitoring on -- In downtime + monitoring on -- In downtime + not monitored - More info on SGM/PRD pool accounts can be found here: https://twiki.cern.ch/twiki/bin/view/LCG/SgmPrdPoolAccounts - The gLite updates pages now have a Change Log to highlight when changes have been made to the notes after they have been released. The Change Log for gLite 3.0 is here: http://glite.web.cern.ch/glite/packages/R3.0/ReleaseNotes_ChangeLog.asp . The Change Log for gLite 3.1 will be created when needed.

more information More_DAG_comments.txt

11:15 → 11:25 Handling tickets

- Please note that Philippa Strange is moving to a new role within SFTC and from September will not be involved with the UKI ROC. Therefore if you have any ticket or other queries that you would usually take to Philippa, please could you now mail the helpdesk directly (ukiroc@grid-support.ac.uk) or take up the matter with your Tier-2 coordinator. Many thanks.

11:25 → 11:30 AOB

- On behalf of everyone in UKI I would like to thank Philippa for her contribution to the ROC over the last few years and wish her well in her new role working on ISSeG. - "Developing and running MPI applications on the EGEE and int.eu.grid infrastructures" 10-11 September Trinity College Dublin, Ireland - See http://www.grid.ie/mpi/wiki/MpiGridUserCourse07.