Minutes for Regional Operations Meeting DECH (June, 30th 2006)

Attendance:
Peter Kunszt (CSCS);
Andreas Gellrich (DESY);
Peter Wienemann, Hans-Gunther Borrmann (Freiburg);
Sven Hermann <chair>, Clemens Koerdt (FZK);
Christian Peter (ITWM);
Andreas Nowack, Thomas Kress (RWTH);
Ute Karabek, Kläre Cassirer (SCAI).

Missing:
Wuppertal (excused),
Dortmund (excused),
MPPMU,
GSI

1. Introduction

Focus of this meeting (biweekly):
for EGEE funded and associated sites within region DECH to improve communication and collaboration with sites. Discuss site specific problems, distribute and collect information easily.

Additional communication channels to be added soon (to avoid discussions via the admin lists.).

#! Sven to clarify existing lists

#! Everybody: Send major topics in advance (to be included in the agenda)

2. Round the Sites (gLite update status, encountered problems, issues to discuss, ops-VO, ..)

*CSCS*
status:    gLite update last week, updating successful, Ops-VO next week
feedback: R-GMA, APEL configuration not straightforward

*DESY*
status:    gLite update was already attacked beginning of June, updating was successful, 3CEs with 250CPUs up and running, some hardware moves have taken place at same time, ops-VO operational
Issues:
- Problems with APEL (YAIM) works only for 1CE/site,
- YAIM FAQ also not helpful in this regard --> Request (YAIM) to provide this in future
- CA-announcements: problems tend to reappear for every new CA update --> Request to bring this forward to the appropriate people

*FZK*
status:    gLite update started on 17.5. and finished on 21.6., ops-VO this week
Issues:
collecting top 5 MW changes: ---> feedback from sites (#! All)

AG: now daily site reports necessary?
SH: no, only one weekly report necessary, but possibility offered for daily changes on it; detailed list of problems not obligatory but can be helpful, more important: give short overview

*GSI*
status: (no site representative present) no gLite update yet (Alice-VO does not want it for the moment)

*ITWM*
status:    update started two weeks ago, updating successful, ops-VO next week
issues:
- usual problems updating of new major releases, same problems tend to reappear,
- too low priority given to many 'hot' fixes

*MPPMU*
status:    (no site representative present) no gLite update yet (only recently certified site), no ops-VO yet

*RWTH Aachen*
status:    no gLite update yet (waiting for template from Quattor working group),    ops-VO operational
issues:
- dCache related SFT-problems (site opened GGUS ticket: #9885),
- Quattor scripts still not available

AG: don't expect to have working scripts from CERN available soon. DESY did quick fix (YAIM component outdated).
SH: fabric management tools like Quattor not officially supported by CERN developers. Apparently, its up to the ROC to provide help with fabric management tools. This was not clear in advance. Better don't wait for complete Quattor scripts from CERN and start upgrading with what you have (DESY can help). ROC will also address this point (again) in the next ROC managers meeting in order to get feedback from other ROCs/Sites (S.H. done).

*SCAI*
status:    current update to gLite 3.0 to be finished by Wednesday 5.7., ops-VO after update
issues:
- R-GMA configuration problematic, status query changes configuration-entries (GGUS Ticket to be opened)
- LFC for dech-VO: permission problems encountered

AG: Problem is that LFC 2.6. had different format compared to new version. Only complete reinstallation solves this. Database structure has changed considerably. Looking for appropriate conversion scripts....
SH: SCAI, DESY, CSCS and Freiburg should work together on LFC issues
PK: CSCS has delegated problem to VOs.

*Uni Freiburg*
status:    updating starts next week, ops-VO together with update

*Uni Karlsruhe*
status:    (no site representative present) no gLite update or ops-VO yet

*Uni Dortmund*
status:    upgrade to gLite 3.0 performed from June 3rd to 6th, OPS VO done since weeks.
issues:
- (more MW and rollout feedback was received by mail),
- problems with dependencies in dCache rpm (GGUS ticket 9204)
- proposal for improvement: clearer guide lines needed, what to install and upgrade

*Uni Wuppertal*
status:    (no site representative present) apparently no mayor problems encountered during update

****

--> !#all:
- please supply ROC with detailed feedback on gLite deployment problems.
- please supply ROC with top 5 MW changes, you would like to have implemented


3. ROC-on-duty (EGEE funded effort)

Status:    Workflow is currently being documented. No documents yet.

Handover CSCS - DESY (P.K.): last week rather quiet with some trailing tickets already addressed.

SH: Please contact your ROC or GGUS for feedback on this.
TK: new support unit for regional CMS related problems in ROC-DECH portal (CMS experts), other support units for different VOs could follow.
SH: Also regional MW Support Unit, announced earlier, for MW related problems.
PK: please use new ROC DECH Wiki (at CSCS) in relation to ROC-DECH-On-Duty

4. Setup of COD-DECH (EGEE funded effort)
Status: Successful training event held on 21.6. with the COD-CERN team. Next week first unofficial COD-shift for ROC-DECH with overall responsibility on Italian COD.

PK: please use CSCS Wiki for this duty as well

5. AOB

AG stressed that there are other VOs then LHC, that are part of the grid and specifically update procedures should respect these VOs' needs. This could also be mentioned at higher levels, as it seems to be a persisting problem in the EGEE/LCG grid (Marcel K, PMB).

-- 
Dr. Sven Hermann                                 sven.hermann@iwr.fzk.de
Forschungszentrum Karlsruhe                        Tel.: +49-7247-828632
Institute for Scientific Computing / Inst. f. Wissenschaftliches Rechnen
Hermann-von-Helmholtz-Platz 1,  76344 Eggenstein-Leopoldshafen,  Germany