PASTE meeting

→ Europe/Zurich
1/1-025 (CERN)

1/1-025

CERN

20
Show room on map
Andrei Tsaregorodtsev
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, 13th November 2007, CERN Chair: A.Tsaregorodtsev Minutes: S. Paterson CERN: M.Bargiotti, G.Castellani, P.Charpentier, J.Closier, E.Lanciotti, S.Paterson, R.Santinelli Phone: R.Bernet, N.Brook, G.Kuznetszov, U.Marconi, M.A.S.Miguelez, R.Nandakumar, A.Tsaregorodtsev, A.Zhelezov Apologies: A.C.Smith 0. Ongoing Action List 23/10/07 - Marianne - Request for basic documentation for restarting the Bookkeeping services 23/10/07 - Marianne - Bookkeeping should be migrated from SLC3 -> SLC4 and installed on volhcb05 23/10/07 - Ricardo - Update Cleaning Agent to prevent user WMS problems 30/10/07 - Marianne - Check number of events for prod 1957 in BK against LFC 30/10/07 - Marianne - Update of LFC replica flags (C checked, N checked but not present) 06/11/07 - Gianluca, Marianne - Remedy the Exceeded Max Staging Retry jobs and update ProcDB as required 06/11/07 - Gianluca, Marianne - Agents for automatic resolution of IntegrityDB entries should be developed 06/11/07 - Raja - Install v2r17p2 on VO-boxes 13/11/07 - Joel - Check dump from Marianne and update ProcessingDB 13/11/07 - Alexei - Produce dump of file namespace at GridKa for Marianne to proceed with LFC replica flags there 13/11/07 - Roberto - Contact site administrators for Tier-1s and install latest version of LFC with Oracle streams enabled 13/11/07 - Roberto - Check status of LFC admin interfaces with WMS instances as trusted hosts 13/11/07 - Roberto - Contact ATLAS representative to understand VOMS usage in other VOs 13/11/07 - Roberto, Gianluca - Resolve issue with Raja's credential causing failed pilot submissions 13/11/07 - Alexei - Investigate recent sporadic failures at GridKa for production jobs 13/11/07 - Joel - To check whether ancestor files are required for reconstruction production 1981 13/11/07 - Roberto - To check the LCG Utilities status for SRM2 at the Tier-1s 13/11/07 - Joel, Matvey - lhcb-proxy-init should be committed and added to the LHCb startup scripts 13/11/07 - Joel - To update VO ID Card to remove obsoleted group definitions 13/11/07 - Marianne - RAL and PIC migrations, FTS job submissions 13/11/07 - Andrei - Preparation of the pilot agent framework document 13/11/07 - DIRAC developers - DIRAC3 (ongoing) Completed Actions Done - 23/10/07 - Marianne - Check dump of ProcessingDB in LFC Done - 23/10/07 - Andrei - Add increased protection for user job deletion in DIRAC2 Done - 06/11/07 - Stuart - StagingSanityAgent should be developed and deployed 1. Production Status Report Minutes from the production meeting are available here. Joel reported that for production 1974, 1975 jobs were taken by Zurich and failed due to a problem with the physics software. Marco has reported a problem at GridKa with loss of connections to data. Alexei is investigating the cause of the failed jobs but reports that no failures are observed for NFS access, investigations are ongoing. ( ACTION: Alexei ) Due to VOMS mapping problems, reconstruction was ok for only 3 Tier-1s but since the last meeting, only NIKHEF still has the issue. The b-inclusive test jobs have been submitted, access to ancestor files seems to be required for these, Joel will check whether this is necessary. ( ACTION: Joel ) Stripping tests have run successfully so these can start to be submitted for low lumi. Andrei and Joel will investigate the disappearance of log files for certain jobs. ( ACTION: Andrei, Joel ) The volhcbYY machines must be rebooted for a kernel upgrade and this will be scheduled for tomorrow morning. This can start with lxgate03 and volhcb03 to check everything is ok. volhcb02 can now be upgraded to SLC4. 2. Bookkeeping Update Marianne and Elisa have started to port the Bookkeeping manager service to DISET. Elisa has been trying to understand how the Bookkeeping works and has started development on the python client for the web page. After the Bookkeeping migration, volhcb01 can be used as a Test system. Philippe summarized the proposal from the Bookkeeping working group. The current functionality of the web page should be replaced with a PyQt interface that can be run in a standalone way. This would allow a better organisation of queries such as obtaining all files generated by a given version of Brunel. Andrei asked how to streamline the BK interfaces for the current implementation and it was suggested that this discussion be postponed until the new developer from Dublin is available. 3. SRM2 Testing Nick commented that some of the Tier-1 sites according to their schedule will deploy SRM2.2 by the end of this month. Discussion with Stuart and Philippe resulted in the idea to ship the upgraded LCG Utilities such that SRM2 can be queried first if available. For uploading output data this becomes difficult because space tokens must be defined in the CS. Philippe queried whether we should change the -tape and -disk definitions to have one for each space. These tests should be performed now with the current production tools in DIRAC2. Latest version of LCG Utilities should be installed at the Tier-1s. There is no mechanism to query the version of LCG Utilities installed at the Tier-1 sites so this should be investigated. ( ACTION: Roberto ) Nick commented that SRM2 tests failed at GridKa due to the required gLite version not being available. Andrei requested to gather all necessary information and discuss at the Thursday operations meeting. 4. Production Operations Issues - Release v2r18 Andrei reported that the job deletion logic is now fixed and secured in v2r18. Packaged with this release will be lhcb-proxy-init that properly initializes the proxy via dirac-proxy-init and can upload the credential to the myproxy repository. Joel commented this could be added to the LHCb startup scripts. Matvey will commit this to CVS. ( ACTION: Joel, Matvey ) - Using VOMS proxies ( roles or groups ) After discussions at the Operations meeting it was agreed to abandon the notion of groups in LHCb and only express use-cases and LHCb user groupings through VOMS roles. Nicknames will be stored as VOMS GAs. The VOMS role should become a parameter of Jobs in the future and the CS will be updated from VOMS regularly. The VO ID Card will be updated today such that redundant groups can be removed. ( ACTION: Joel ) Roberto commented that there are different possibilities at registration time from the VO Manager to insert new users. VOMRS will be able to propagate changes of DN to the LFC via the immutable 'nickname' field but this is not currently available and should be investigated. Andrei remarked that an LFC proxy admin service can be exposed to Joel as VO manager to perform and automate certain tasks. - RAL and PIC data migration Marianne reported no updates since yesterday, plans were added to a Twiki page. FTS job submission is ongoing. PIC migration is scheduled in two weeks from now. The LFC people are not yet informed of the migration because the dates are not yet fixed. ( ACTION: Marianne ) Raja commented that the migration timescale would be 10 days assuming the FTS jobs proceed without delay. 5. Pilot Agents Usage Andrei attached a draft document to the Agenda page. This clearly highlights the mechanism within DIRAC for this and comments are welcome. Stuart commented that it may be worth explaining how the generic pilot proxy could be secured in the glexec logging mode. This is not an issue for the set UID mode. ( ACTION: Andrei ) 6. DIRAC3 - Workload Management status (job submission, optimizers, monitoring) Andrei mentioned that basic monitoring and submission tools are available. Stuart commented that progress had been made with the optimizers. A simple API version is being prepared now in order to test the full chain. The monitoring webpage is available in a preliminary status and Matvey is continuing with this in the new web framework. - Data Management status The LFC client has been finalized and ported to DIRAC3. Prototypes of the Storage classes for e.g. SRM2 are available. The Storage Element class will be heavily reimplemented (originally a DIRAC2 Review recommendation). After this the Replica Manager will be started. - Production Management status Gennady is working on a Repository Database and is prototyping the commands for submission of production workflows. The API will still be used to submit the jobs to the WMS. - Release tools Tools for making releases starts to become urgent when testing job execution and submission etc. Joel (and Ricardo when available) will try to proceed with this. 7. AOB Philippe proposed a table for DIRAC3 development goals and proposed timescale, Andrei agreed to prepare this.
There are minutes attached to this event. Show them.
    • 09:30 → 09:45
      Production status report 15m
      - Report from the Production meeting:
    • 09:45 → 10:15
      Production operations issues 30m
      - release v2r18 - Update on the Staging strategy, data integrity checks - ProcDB checks and updates - RAL and PIC data migration - Using VOMS proxies ( roles or groups ) - lxgate34 usage
    • 10:15 → 10:35
      Pilot agents usage 20m
      document
    • 10:35 → 11:05
      DIRAC3 tasks 30m
      - Workload Management status (job submission, optimizers, monitoring) - Data Management status - Production Management status - Bookkeeping status - Release tools
    • 11:05 → 11:15
      AOB 10m