Choose timezone
Your profile timezone:
PASTE meeting
→
Europe/Zurich
Description
CERN, bat 2 R-030
To join the telephone conference, click here: https://audioconf.cern.ch/call/0122882
To dial in to the conference:
a. Dial +41 22 7676000
b. Enter access code 0122882
Production and PASTE Meeting, 29th May 2007, CERN
Chair: A.Tsaregorodtsev
Minutes: S. Paterson
CERN: M.Bargiotti, G.Castellani, P.Charpentier, J.Closier, A.Smith, S.Paterson, R.Santinelli
Phone: R.Bernet, D.Bortolotti, N.Brook, M.Sapunov, A.Tsaregorodtsev
1. Production Status Report
Minutes from the production meeting are available here. Joel reported that only reconstruction was running in the
previous week.
Below is a summary of the issues by site:
- At PIC there were problems with rfio at certain times, resulting in job failures ( ACTION: Ricardo )
- RAL has performed very well with between 1 and 2.4K successful stage requests per day
- GRIDKA has been running some jobs without problems but there have been a significant number of stalled jobs
- Over the previous weekend only 18 jobs ran at CERN due to tapes blockages ( ACTION: Joel )
- Few jobs are remaining for NIKHEF / IN2P3 / CNAF.
For the issue of rescheduled jobs Philippe mentioned that the Job Wrapper should check if outputs are already present
before consuming CPU resources. Stuart agreed to put this in place for the DIRAC3 Job Wrapper. With DIRAC2 it has
always been difficult to distinguish between successful jobs in the stalled status and those which really failed. Stuart
suggested a WMS agent to distinguish between the two cases.
Joel conveyed that the latest Gauss version is being tested with new workflows. The requests have been received from
Olivier and simulation will restart after checking the event types are correct. Joel also reported on the latest
difficulties with the new CERN certificates, at NIKHEF and LYON the new CA is not recognised and gsidcap data access
fails. There is an ongoing exchange of mails with the site representatives to resolve this as quickly as possible. Joel
agreed to follow this issue. ( ACTION: Joel )
It was agreed that the tests for local data access should be added to the SRM critical tests for LHCb (although using
the default LHCb VOMS mapping not SGM).
All pending requests were served by the VO-boxes for reconstruction jobs.
2. Production Operations Issues
There was some discussion about the use of root protocol at CERN. Stuart mentioned that recent tests were fine at
PIC and CNAF (excluding the DC04 DaVinci failures with RFIO). At CERN there seems to be a problematic WN setting
and Philippe agreed to contact Jan about this. ( ACTION: Philippe )
Philippe mentioned that the Root GFAL plugin status is unchanged. The issue of determining currently running
software on Grid WNs is still unresolved and Philippe will raise the point at the next Architects Forum. Regression
tests should be available for the Root team to experiment with various implementations of dCache and DPM running
the same software versions. ( ACTION: Philippe )
A mechanism should be made available to Ganga to support the LHCb packaged versions of Root via the Applications
Area. Stuart agreed to investigate this with Marco and Hubert ( ACTION: Stuart )
Marianne reported on the DC06 cleanup status, att IN2P3 all files were cleared. There is still no response from the
LFC support regarding the bulk removal of DIGI and SIM files (now > 1 month waiting). Marianne mentioned that
the LFC had now been cleared and the Tier-1 disk storage cleanup is underway. 200 files had recently been lost at
PIC, the deletion will be postponed until Carmine returns in order to determine the GUIDs of the files.
There was some discussion of site by site splitting of jobs from Ganga and no news had been forthcoming on this.
Stuart agreed to investigate the status. ( ACTION: Stuart )
Joel mentioned that the SRM durable endpoints at CNAF and CERN were causing problems with the POOL XML
schema. This should be investigated and understood. ( ACTION: Joel )
Marianne is preparing the lists for the SA path cleanup and mentioned that CERN_Castor had been removed from the
LFC thanks to a script from Roberto.
Roberto raised the issue of SAM SRM critical tests, there was no Operations meeting this week but the question of
whether this test should be run as operations was asked. After some discussion it was agreed these should be SRM
not LHCb specific tests, since problems with SRM endpoints reflect Grid not LHCb issues. Roberto was asked to convey
this as an LHCb request at the Operations meeting and Joel will run LHCb tests in parallel. ( ACTION: Joel, Roberto )
Andrei reported that Matvey had run successful simulation tests on NorduGrid and a higher level discussion may be
required to proceed with further use of the resource. With Windows resources and Irish sites also cropping up Andrei
wanted to make the point that LHCb is short on CPU and the policy for pursuing these resources should be made
clear. It was suggested that Andrei contact Torre Wenaus regarding tests on OSG ( ACTION: Andrei )
3. DIRAC3 Tasks
In the absence of Adri and Ricardo, Andrei mentioned that the DIRAC3 framework is progressing rapidly. There have
been many fixes and the introduction of the CS client is complete.
Joel reported there has been no progress with the deployment tools due to operational issues. Gennady has
successfully started a DIRAC3 service and work is ongoing. Andrei is currently testing clients. Andrew has started
formulating a generalised Data Managment request and the streamlining of the DM components is underway. Stuart
agreed to get started on the Job Wrapper this week.
4. AOB
Philippe proposed a Dataset discussion meeting for the 5th June during the LHCb week. The reason for this is to
ensure a consistent dataset concept for LHCb. Marianne was tasked with presenting the DIRAC point of view at the
meeting ( ACTION: Marianne )
Andrei proposed that the next DIRAC3 week will be during 25th-29th June.
There are minutes attached to this event.
Show them.