CCRC08DailyMeetingsWeek080310 < LCG

LCG Web>WLCGCommonComputingReadinessChallenges>CCRC08DailyMeetingsWeek080310 (2008-03-14, HarryRenshall) (raw view)
-- Main.HarryRenshall - 07 Mar 2008
---+ Week of 080310
Open Actions from last week: Respond to RAL what are the LHC experiment requirements for local RB and WMS at RAL.


---++ Monday:

See the [[https://twiki.cern.ch/twiki/bin/view/EGEE/WlcgOsgEgeeOpsMeetingMinutes][weekly joint operations meeting minutes]]

Additional Material:

ATLAS: Finished M6 this morning. Went well with 40 TB data over the weekend and calibration streams going to the 4 T2 as planned (Rome, Naples, Munich, Michigan). This week functional tests of T1 to T1 between Lyon and PIC then next week throughput tests primarily of T0 to T1.

ALICE: Continuing data export to 5 working sites and planning to add RAL this week.

CMS: See detailed report in EGEE minutes. Following up on LSF failure at CERN which blocked T0 operations. Reprocessing going well with some site issues at FNAL and IN2P3. In T1 to T1 commissioning only RAL to ASGC is still missing.

LHCb: Preparing for stripping jobs. Setting up a way of reporting remaining CP time for grid jobs. Setting up srmv2 endpoint tests to go into LHCb SAM.

PIC: Downtime next week 17-20 March.

FZK: Oracle updates Sat 15 Match will put LFC 'at risk'.

---++ Tuesday:

elog review: One new entry Tuesday morning- srm put failures to CERN LHCb durable disk. Problem found and fixed soon after.

Experiment report(s): 

LHCb (RS): Since some time LHCb has been migrating data at RAL from dcache to castor and it is about 50% complete. They now want to speed up this procedure by transferring files that are also at CERN directly from CERN leaving only the files uniquely at RAL to be recalled from dcache. The LFC mirror at FZK is now operational.

ATLAS (AdG): The upgrade to DQ2 version 0.6 went well. This week they will be testing T1 to T1 transfers to PIC and IN2P3. They are preparing a sample to be sent to each T1 of 20 datasets of 100 files each as the test data. Each two weeks they will exercise different T1 sites as the target. In the alternate weeks they plan T0 to T1 throughput tests which include also exercising T1 to T2 transfers. This pattern will continue up to the May CCRC'08. 

CMS (DB): Focussing on CERN to T1 transfers but this will reduce during this week as they will move to debugging links. They would like to coordinate some throughput tests, including T1 to T1, with ATLAS during March/April. Their T0 processing has been suffering from some job wrapper errors.

ALICE (HR): Continuing high rate export to their 5 T1. RAL is working on their ALICE setup but at low priority. 

Core services (CERN) report:

DB services (CERN) report:
 
Monitoring / dashboard report: 

Release update: 

Questions/comments from sites/experiments: M.Jouvain at GRIF reported a problem of ATLAS transfers from IN2P3 requesting the SRMv1 transfer protocol but having SRMv2 space tokens.

AOB: 

---++ Wednesday

elog review: Nothing new

Experiment report(s):

ATLAS (AdG): Have started T1 to T1 transfers with PIC as the first target with, initially, functional tests to look for any particular problems. The PIC VObox has been upgraded to DQ2 0.6 for this purpose.

CMS (DB): They have started a 'global run' which is Tier 0 only and have some setup issues with Castor service class environment variables. T0 to T1 exports are ramping down to the 50% level to allow sites to clean up disks and tapes.

LHCb (RS): Continuing preparations for the May run. Some LHCb LFC are not yet being published in the IS. They would like access to a prototype of gLEXEC. Next week is a software week which will include, on Wednesday afternoon, reports on the February CCRC'08 run.
 
Core services (CERN) report (GM): The FTS team will be restarting some dteam export in order to test various patches. Initially this will be low rate and any stress tests will be first scheduled through this meeting.

DB services (CERN) report (MG): There was a problem on Monday during a scheduled move of a Gridview table which made Gridview unavailable for some time and also prevented the ATLAS SAM tests from running. The cause was a lack of space in the production database and a meeting to review this, e.g. how long to keep data, has been arranged for next Monday. 

Monitoring / dashboard report: The SAM problem of recognising both the new and deprecated srm publishing syntax is now being addressed with a new script in preparation. 

Release update: 

Questions/comments from sites/experiments: NT asked what future gLite releases might be needed in the next weeks. GM said they may have a new FTS and at some time the Oracle version of AMGA will come out of certification and testing. 

AOB: 

---++ Thursday

elog review: The bdii2oracle.py script has been fixed so SRM instances (both the new and deprecated syntax) are all visible again.


Experiment report(s):

ATLAS: Debugging T1 transfers to PIC where rates are not as high as expected and many transfers are stuck. The PIC configuration needs tuning but is shared with other VOs so must be negotiated. 

CMS: Still having trouble with getting their global run data visible to the CERN CAF.

Core services (CERN) report:

DB services (CERN) report:

Monitoring / dashboard report: 

Release update: 

Questions/comments from sites/experiments: 

AOB: 

---++ Friday

elog review: LHCb reported unset ACLs on some of their CERN Castor spaces - now fixed.

Experiment report(s):

ATLAS: Have had at least 3 problems in the T1 tests to PIC and they will soon be going down for several days so ATLAS will try to debug T1 to IN2P3 now. Should be a cleaner situation as they already have DQ2 0.6 deployed. They will continue debugging next week and postpone the Tier 0 to Tier 1 throughput tests till after Easter.

CMS: They are deciding how to assign Tier 2 sites to analysis groups and preparing for cms site operations relying on the SAM tests.

Core services (CERN) report:

DB services (CERN) report: Both PIC and CNAF will be having long down times so the splitting off of their streams replication has been prepared and they have optimised the window to accomodate these long downtimes. The new RAC hardware has been installed and Oracle installations are scheduled for next week.

Monitoring / dashboard report: 

Release update: 

Questions/comments from sites/experiments: 

AOB: 

---++
Topic revision: r4 - 2008-03-14 - HarryRenshall
LCG Wikis
LCG Service Coordination
LCG Grid Deployment
LCG Applications Area
Public webs
Welcome Guest
- Cern Search
- TWiki Search
- Google Search
LCG All webs