UKI Monthly Operations Meeting (TB-SUPPORT)

Thursday 21 Feb 2008, 10:30 → 11:30 Europe/Zurich

EVO - GridPP Deployment team meeting (under GridPP community)

Monthly review and discussion meeting for those involved with UKI EGEE/WCG deployment and operations. The phone bridge number is +41 22 76 71400. The phone bridge ID is 239486 with code: 4880.

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

- Accounting -- Accounting enforcement is stepping up. If a site has not published for >31 days the *critical* publishing check test (SAM- apel-publish) will fail affecting site reliability and availability figures -- Please use this link to check the status being reported to EGEE: http://www3.egee.cesga.es/acctenfor/nodata.php - Monitoring status -- SAM tests --- http://hepwww.ph.qmul.ac.uk/~lloyd/gridpp/samtest.html (for GridPP). Average is still below target. The charts (click on an entry for your site http://hepwww.ph.qmul.ac.uk/~lloyd/gridpp/samplots.html) will be placed in the wiki area for comment shortly. --- SL ATLAS tests are still being used http://hepwww.ph.qmul.ac.uk/~lloyd/gridpp/atest.html. The success history has been archived for each site and we will produce plots for comment over the coming months. - Site concerns -- Your chance to raise problems and issues

10:50 → 11:05 Current experiment/VO activities and issues

LHC experiments have been working on the Common Computing Readiness Challenge for the last 2 weeks. - ATLAS -- Need SL4 WNs -- Need storage tokens setup at sites. This has been the biggest T2 issue for them (Graeme has added an item on DPM and how to set up ATLAS space tokens here: http://scotgrid.blogspot.com/). -- Glasgow had problem with FDR data which will affect all sites on DPM which concerns default write permissions - not group writeable as needed by shifters. Solution is additional ACL fix. -- After upgrades and when space tokens implemented sites *must* publish their endpoints to be used! -- Next week random data transfers are expected to stress the system -- Production wise more UK sites now validated. Recovering QMUL files being pursued as otherwise the production system acts as if more jobs to run and wastes time. -- Site status information is being tracked here: http://www.gridpp.ac.uk/wiki/AtlasFdr1 - CMS -- UK wise recovering well after T1 power failure and subsequent downtime -- Managing 100MB/s (T0-T1 targets ~40MB/s). -- T1-T2 targets 20MB/s and getting 25-30MB/s (but still lower than tests 2 yrs ago) -- Main GridPP sites thought to be ready -- There is a problem with proxies getting corrupted - workarounds in place. - LHCb -- CCRC full tests running, no major problems for T0-T1. -- Struggling with reconstruction at some T1s - RAL is ok. -- Old lcg-utils and sites on SL3 (WNs) now causing problems. Sites need to be at 1.6.7. SL4 WNs are now critical as latest software is not being built on SL3. ATLAS have a software week coming up http://indico.cern.ch/conferenceDisplay.py?confId=22132. CCRC will be reviewed on the Wednesday. There is also an overall pre-GDB review of the first part of CCRC on 4th March: http://indico.cern.ch/conferenceDisplay.py?confId=29170. - Regional VOs -- ScotGrid now has VO enabled for users. LT2 running for some time. What of the others? -- supernemo *still* asking for more sites to support it in the UK. Please support it! - Others -- In process of following up issues surrounding many sites seeing inefficient biomed jobs. The VO manager is aware of this problem and is going to encourage users to respond.

11:05 → 11:20 Recent PMB/GDB discussions

GridPP PMB *************** 1) Tier-1 power outage raised the issue of disaster planning and management. Have a look at the new Tier-1 blog for a short summary: http://www.gridpp.rl.ac.uk/blog/ [Are there specific things that T2s would like to see in this Blog?]. 2) Would like a transition plan for the move from LCG-RBs to gLite-WMS. We are currently waiting for the SL4 version of the WMS. There are SL3 versions at RAL, Glasgow and IC. camont have reported a factor of 6 improvement for job throughput. We need to recirculate links to the user documentation (it is in the usual places...) 3) Defining the GridPP structure for GridPP3. WLCG GDB ***************** Last meeting was on 6th February: http://indico.cern.ch/conferenceDisplay.py?confId=20226 - Authorisation - Job priority progress: Various (PBS) installs on PPS. Testing queues and VOVIEW information. Checking behaviour with different VOMS FQANs. Things worked but issues with ease of configuration. - 64-bit resources - Traceability and Logging Policy: defines minimum requirements - Accounting updates: new APEL features such as user based views. Portal reports. Few UKI sites publishing FQANs, more publishing UserDNs (publishGlobalUserName needs to be set to yes if site wants to publish. YAIM setting coming in form of APEL_PUBLISH_USER_DN - turns encryption of user DN on). - Storage accounting. Greig talked about the storage accounting portal. - Pilot jobs: revised policy https://edms.cern.ch/document/855383/1. Now takes account of traceability and logging requirements - please read and comment! Glexec remains the only candidate for job ID change. Glexec review team now in place and has a mandate from MB. - Tape efficiency - CCRC - Daily 15:00 (CET) meetings. Proposal for WLCG T2 coordinators. Status views: http://gridmap.cern.ch/ccrc08/. Sites encouraged to complete e-logbooks: https://prod-grid-logger.cern.ch/elog/CCRC'08+Observations/. See also: http://prod-grid-logger.cern.ch/feeds/CCRC%2708+Logbook/elog.rdf. UK T2s please use the GridPP blogs. - Resource Scrutiny Group Next GDB is on 6th March: http://indico.cern.ch/conferenceDisplay.py?confId=20227

11:20 → 11:25 ROC/ops notes & issues

- The GGUS to Footprints importer stopped working after a Footprints configuration change - For the time being a manual interface is in place. T2s encouraged to watch for UKI tickets in GGUS. - 64-bit WN release has just gone to the PPS - YAIM developers introducing error messages and exit codes. They want your input on how to implement this for various configuration management systems. - The classic SE is being phased out by May. To migrate see https://twiki.cern.ch/twiki/bin/view/LCG/ClassicSeToDpm. - The 3rd EGEE User Forum took place last week. If you have interested groups at your site then they may be interested to first look at the forum summary talks here http://tinyurl.com/27mpx3 and here http://tinyurl.com/28f9h8 which cover the following areas: Fusion Computational chemistry and material science Earth science Finance & multimedia Astronomy & astrophysics Life sciences Monitoring, accounting & support Data management Grid access Interoperability and resource utilisation Workflow and parallelism

11:25 → 11:30 AOB

- A second MPI workshop is being planned at CNAF, Bologna on March 19th and 20th.