LCG Management Board

 

Date/Time:

Tuesday 28 February 2006 at 16:00

Agenda:

http://agenda.cern.ch/fullAgenda.php?ida=a057124

Members:

http://lcg.web.cern.ch/LCG/Boards/MB/mb-members.html

 

(Version 2 - 5.3.2006)

Participants:

A.Aimar (notes), D.Barberis, L.Bauerdick, I.Bird, K.Bos, Ph.Charpentier, D.Foster, B.Gibbard, J.Gordon, E.Laure, H.Marten, G.Merino, B.Panzer, L.Robertson (chair), Y.Schutz

Action List

https://twiki.cern.ch/twiki/bin/view/LCG/MbActionList

Next Meeting:

Tuesday 7 March 2006 from 16:00 to 1800

1. Minutes and Matters Arising (minutes)

 

1.1 Minutes

No feedback received. Minutes of the previous meeting approved.

 

1.2 Face-to-face MB Meeting in Rome: 4 April 2006 at 16:00

The meeting will be held in CASPUR, Rome, on the day before the GDB.

The phone conference setup will be the usual one.

 

1.3 GridPP Resource Analysis (document)

The document attached explains how Tier-1 and Tier-2 resources have been “promised/delivered/available/used” to support GridPP activities in the UK.

This is a clear and succinct summary on how resources are used. It could be a good example for the kind of reports about resources and accounting that LCG could provide in the future.

1.4 Internal Services Review - dates

The dates agreed are the 8-9 June 2006, during the week of the GDB meeting (7 June) and preceding the Tier-2 Workshop (12 June).

2. Action List Review (list of actions )

 

         25 Feb 06 – MB members and SC4 Contacts should send feedback and clarification to J.Shiers on the document about “Conclusions, Decisions and Open Issues after the Mumbai SC4 Workshop”.

 

Feedback received and the updated document is discussed during this meeting.

 

         28 Feb 06 - Experiments should clarify with B.Panzer the exact schedule (week no) and resources needed (KSI2000, etc.) in 2006.

 

On the way. A proposal with the available information will be presented next week during the MB face-to-face meeting.

 

3. Quarterly Report Review – A.Aimar (documents)

 

The Quarterly Reports Summary and a short Review have been distributed to the MB and to the authors of the reports in order to allow for feedback to improve the documents. Among other improvements that are needed the authors were requested, in particular, to complete in a more textual format the “Summary of Progress” (couple of paragraphs would be sufficient) so that this text could be part of a single Executive Summary for the Overview Board.

 

Some members of the MB felt that the Summary of Progress should not necessarily be in a textual format that could be integrated into the Executive Summary, and so this request was dropped. The Summary of Progress field will not be used in the Executive Summary for the Overview Board.

 

Once the reports are final and submitted to the Overview Board, the “milestones changes” or “new milestones” that are specified in the reports will go through a change review and the plans of each project will be updated.

 

The request for associating each completed milestone with some deliverable or document confirming the completion of the milestone is not always applicable. Therefore it is not mandatory but should be specified when possible.

 

4. Feedback on the Mumbai Workshop Conclusions – L.Robertson (document)

 

 

The updated document takes into account the feedback received with the explanation of what was taken into account or discarded.

 

The intention is to have the document approved by the GDB on Wednesday next week. Therefore feedback from GDB members is expected by Monday 6 March t 13:00. A.Aimar will send the document to the GDB on behalf of K.Bos (traveling) in order to receive GDB feedback in time.

 

Two issues that are clarified in the introduction to the updated document concern XROOTD and VOMS.

4.1 Use of XROOTD by ALICE (Section 2.2 in the document)

 

In SC4 ALICE will continue to use xrootd in their VO Boxes and to discuss directly with the sites about deployment and installation.

 

The integration of xrootd and the mass storage systems is a longer term issue. A meeting is being organized at CERN with Andy Hanuchevsky and mass storage experts to discuss and compare the possible solutions for interfacing xrootd to the mass storage systems. It is hoped to organize this meeting during March.

 

4.2 VOMS with group and role support in SC4
The VOMS version that supports groups and roles will be available in SC4, but the only package that uses this in SC4 will be the DPM. Availability of the support in VOMS will allow other SRM/MSS implementations to be developed and deployed, but this is not in the SC4 plan. It will also enable implementation of scheduling policies using groups and roles, but the way in which this will be done has not yet been agreed.

 

Kors Bos noted that at the last GDB Cal Loomis was asked to investigate the experiments needs in this area. Erwin Laure said that this issue is also being tackled by the EGEE TCG through a working group led by Dietrich Lico. This concentrates on support by the EGEE WMS, not to other scheduling systems.

 

4.3 Other issues discussed

 

The CE deployed in SC4 will be the implementation that includes Condor C, which has been under test for some time, including availability on the Pre Production Service. Other CE implementations, such as CREAM, will be considered for deployment after they have gone through the full process of evaluation and testing.

 

 

5. gLite 3.0 Status – I.Bird (transparencies)

 

The attached transparencies on the agenda provide a summary of the Glite 3.0 status.

 

The schedule is unchanged:

-          End February until 15 March: gLite 3.0 to be deployed on the PPS.

-          From March to End April beta testing by the experiments.

-          During May deployment in production

-          1st June production starts

 

As agreed the goal is to provide gLite 3.0 as LCG 2.7 enhanced by components from the gLite 1.5 code release. The gLite 1.5 release was made on Jan 21, but there were many problems with this release (see slides no 2 and 3) which have taken some time to fix. Too many components had not been tested sufficiently before the release, including the most important new component, WMS, where the 1.5 version had many problems not present in the 1.4.1 version that has been available on the pre production service. Because of these problems it has not yet been possible to run any stress or performance tests. These tests will now have to be done in parallel with the beta testing.

 

In summary there have been many difficulties, but they are being recovered and the plan is still to keep to the original timescale.

 

In order to define correct priorities in case of further problems for the WMS Ian asked for clarification of the experiments requirements. LHCb is concerned by the submission time needed – this is not a problem for production, but will be for analysis work. Bulk job submission is a priority for all experiments.

 

Ian Bird said that in future the procedure to prepare a distribution of the middleware will have to change:

-          Deployment will be done component by component, no longer coupling components into general code releases.

-          Development must be done according to clearly agreed priorities, avoiding additional features added by developers without agreement.

-          Components will only be accepted for deployment when they have been shown to be able to be installed and pass basic tests.

-          A permanent pre-production service will be operated, including new components that have passed certification.

-          Components will only be included in the distribution when they have been thoroughly tested on the pre-production service.

 

During SC4, if a severe problem will have to be fixed and clients need to be redistributed to the experiments there will be a case by case analysis of the best way to upgrade to the new version.

 

6. AOB

 

6.1 Next meeting with the LHCC referees - 21 March

 

The meeting will be on Tuesday at lunchtime.

 

The topics to cover are:

-          Update on SC4 plans

-          Progress of the 3D deployment

-          Status of SEAL and ROOT migration.

 

Representatives from the experiments and from some sites will be required to be present. To be discussed at the next MB.

 

6.2 Schedule for agreeing on SC4 resource usage at T0, T1s

 

Not discussed at the meeting.

 

6.3. Next MB meeting

Next MB meeting will be a face-to-face meeting at CERN.

 

6.4 Review of the LCG TDR

A report was made available to the LHCC. This has not been seen by the LCG management, although it is dated Nov 2005. L.Robertson will ask permission to distribute it to the MB.

 

7. Summary of New Actions

 

 

No new actions assigned.

 

The full Action List, current and past items, will be in this wiki page before next MB meeting.