Deployment team

Tuesday 5 Feb 2008, 11:00 → 12:15 Europe/Zurich

EVO - GridPP Deployment team meeting

Jeremy Coles

- This is the weekly DTEAM meeting - The intention is to run the meeting in EVO: http://evo.caltech.edu/evoGate/. Join the meeting in the GridPP Community area - The phone bridge number is +41 22 76 71400. Th phone bridge ID is 124333 with code: 4880.

Present ------- Andrew Elwell Derek Ross Pete Gronbech Duncan Rand Raja Nandakumar David Colling Alessandara Forti Graeme Stewart Fredric Brochu Jens Jensen EXPERIMENT ROUNDUP ------------------ LHCb - preparing to ramp up for ccrc08 ral fine for castor - some probs with SAM tests (pilots / RB issue - Cern investigating -- CMS - similar - ccrc rampup. hasn't really started yesterday. InterT1 transfers for UKL and T1-T2 should be getting 45MBs from Fermi next week T0 T1 transfers castor - will see how its going tomorrow. - seems OK at the moment. -- ATLAS - prodn gone OK in last week UK running 2-3k jobs/day glitch over weekend as problems with atlS ORACLE AT CERN executors unable to pick up jobs (nos dropped in UK - now going back up GS validating other T2s UCL-HEP now in. LESC comining post LFC client upgrade - will be a Q on IC-HEP as they don't have their own storage. Shef - pilot job issue. Camb - condor wont run jobs. FDR supposed rto start yesterday (data from pit -> whole chain) - held up as the mixing took longer than planned has now started. pit->castor OK. 1st histograms coming Real Soon Now. Data @T0 today, T1tomorrow. for FDR *are* happy to use SRMv1 but for CCRC want SRM2.2 - only 3 T2s have space tokens for atlas. if you're running 1.6.5 then easy upgrade to 1.6.7 and from there there's reacipies for getting the info provider. (documented) -- Other exp issues - none. --- ROC update ca update - why can't SAM tests detect 2 versions during changeover. with egee3 there'll be a share of runninig CIC On Duty bit. GDB tomorrow. CCRC F2F today. SL5 / 64bit * Which sites are running 64bit SL4 - Consensos 'most sites' - Glasgow, Impeial, UCL, Lanc Other Northgrid sites are 32. * send a summary to dteam mailing list. if they producxe the 64 bit middleware, sites will instakll it. Sl5 - Dave C has virtual SL4 machines within SL5 but thats the only known. QM 'interested' but not using it yet. External pressures - ecdf systems want it in ~6-12 months? need m/w available. Will cern be ready with SL5? apparently it's 'easy' with etics but no sign of any packages... CCRC08 readiness ---------------- experiments - See above ATLAS - SPACE TOKENS!! NOW!! LHCb pit to T0 test should have started. Which have upgraded - Southgrid 'intending' and ready 'quite soon' Glasgow + ECDF ready. Durham - bit messy - couldn't get any DN's for creating gridmap file - PG also saw same issue. - this machine was running an uncertified 1.6.6 64but version. Same as PG. Northgrid - liv will upgrade this week (they are down just now) - manchester in progress. Storage systems should be ready for atlas data next week. MCR also had a large range of certs about to expire - renewal in progress. Tickets and Actions ------------------- footprints ---------- 2181 - GS discussed with jeremy - can be closed. No prodn at shef right now so not sure if it is an endemic problem. 2343 - Alessandara - Perhaps lancs PNFS database? should have been solved by now. ... Proposal to check after meeting and close any old ones - accepted. Footprints - can't close tickets unless they come from the original mail message (ie, not the generic site email addresses) see Footprints 2610 for GS list of complaints. dteam actions ------------- reviewed - D-071009-1 - closed.AE mail Derek directly for info D-071015-1 - Closed. Done. D-071106-1 - Closed - installed and running D-071120-1 - Done - 'In Progress' on remote ticket. D-071120-2 - ongoing. GS still to do. D-080115-1 - Alessandara sent comments to Jeremy - documented way on wiki is easier. - close. AOB --- Jeremy to clarify GridPP3 membership GridPP20 - Ireland = overseas - get the RAL forms off. WLCG workshop - Sysadmins + experts funded by GridPP Tier1 discussions - need a worked example from jeremy. RAL has new disk servers - Contact martin Bly for Procurement details / info. Viglen disk servers - unsure about CPU. Alessandara - another ticket raised for biomed - Contacting a non-responding server in Israel for 'pilots' - killed jobs and banned user. Duncan also saw an italian name doing the same. some biomed users using BOINC and their pilot was checking the lifetime of the proxy and grabbing more. Several other ROCs have ticketed biomed. Anyone going to EGEE User forum - Yes - Jens with a demo and talk. Katie, Claire (NGS) and Santanu from Cambs. 2 from Glasgow - Morag Burgon-Lyon and an (possibly 2) engineers. Getting rid of verified tickets in footprints - Has to be done by Steven. Wrap up at 12:02.

11:00 → 11:20 Experiment problems/issues

- LHCb - CMS - ATLAS - Other

11:20 → 11:35 ROC update

- ROC manager's meeting this afternoon. The agenda is here: http://tinyurl.com/2m74tb. Main things up for discussion are accounting; inappropriate jobs; BDII-GOCDB mismatches; survey of sites for GGUS requirements(?); the CA update procedure and making CE-sft-posix a critical test. - There is a COD meeting following the All ROC Manager's (ARM) meeting. The agenda is here: http://indico.cern.ch/conferenceDisplay.py?confId=21867. This is going to become increasingly important for T2s as the UK COD work needs to be distributed across the region in EGEE-III. Topics on the agenda are: feedback on dashboard and operations manual; tools; failover procedures and SAM integration. - There is a GDB meeting this Wednesday with a CCRC F2F today (Tuesday): http://indico.cern.ch/conferenceDisplay.py?confId=26922. The GDB agenda is here: http://indico.cern.ch/conferenceDisplay.py?confId=20226. The bit on 64-bit resources can be preceded with a discussion at DTEAM this meeting. Most of the other areas are updates or discussions we've seen before - though note there is something on traceability and logging. - EGEE ops meeting -- No ROC items this week -- UKI is backup COD this week -- Do we have any comments from being TPM last week? -- CCRC started Monday 4th. ATLAS will not have major data export until at least Wednesday.

11:35 → 11:45 64-bit resources (and SL5)

- Which sites are running with 64-bit services or WNs? - What are the main issues being faced - What is the GridPP comment on barriers to deployment - do we feel it is worth doing? - I've seen one question about moving to SL5. Is there any immediate widespread concern/need to start raising concerns about SL5 readiness?

11:45 → 11:55 Readiness for CCRC'08

- Status from the experiment side - Which sites have upgraded and which show no signs of moving to the baseline state - What are the experiment timelines and what is planned - Which sites *will* be used according to their current readiness? - Do we need to focus on any particular sites/issues?

11:55 → 12:05 Ticket & Actions review

- Please see the linked page for tickets - Please refer to the actions list: http://www.gridpp.ac.uk/wiki/Deployment_Team_Action_items

more information report_file20698.html

12:05 → 12:10 AOB

- From last week's F2F PMB meeting Jeremy has actions to: -- Clarify DTEAM membership in GridPP3 (who is expected to attend) -- Ensure that the T2Cs are aware that more of the COD task must be taken up by the T2s/UKI EGEE partners in the coming months - plans to be discussed at next DTEAM meeting -- Remind sites that a visit notice is required for GridPP20 (also remind people to register!) -- Inform sites (also for UKI meeting next week) that the PMB will once again allow one sysadmin from each site (plus experts) to request GridPP funding to attend -- Gather input on Tier-1 areas to discuss at slots in future DTEAM and UKI meetings