LCG Management Board

 

Date/Time:

Tuesday 15 November 2005 at 17:00

Agenda:

http://agenda.cern.ch/fullAgenda.php?ida=a057091

Members:

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

 

(Version 1 - 16.11.2005)

Participants:

A.Aimar (notes), D.Barberis, L.Bauerdick, L.Betev, I.Bird, F.Hernandez, N.Brook, T.Cass, Ph.Charpentier, Di Quing, I.Fisk, J.Gordon, B.Gibbard, M.Lamanna, H.Marten, P.Mato, M.Mazzucato, A.Van Rijn, G.Merino, J.Shiers, L.Robertson (chair)

Next Meeting:

Tuesday 22 November 2005 at 1600

Minutes and matters arising (previous minutes)

 

The minutes of the previous week were only distributed in the morning.
MB members can send comments during the week.

 

Actions (list of actions)

 

  • Tier-1 Sites (at GDB) should complete their plans following the guidelines of the presentation given at the GDB. (GDB presentation).

 

The plans should be sent to the PO by the 16 Nov 2005 (see contact on wiki page) .

 

  • ATLAS will send a proposal about reporting on issues concerning non-EGEE resources to the MB.

 

Not done yet.

 

  • CMS will send a similar statement to formally confirm the CMS opinion on OSG reporting.

 

Not done yet.

 

  • L.Robertson will announce the decision to stop the Phase 2 Planning Group..

 

Done (after the MB meeting).

 

  • Tier-1 Sites to set milestones on their known Tier-1-Tier-1 and Tier-1-Tier2 testing activities (IN2P3 and RAL stated specifically that they will do this).

 

RAL done. IN2P3 not done yet because will be discussed in the next few weeks.

 

  • All MB members - Preparation for the Mumbai workshop – details should be agreed following the 9 November GDB.

 

The MB discussed in more detail the preparation of the workshop during this meeting.

 

  • All MB members to provide more feedback on the High Level Milestone is needed.

 

If no feedback comes next week, the High Level Milestones will be distributed for formal approval at next MB meeting.

 

Action: L.Robertson and A.Aimar will prepare the High Level Milestones for formal approval by the MB.

 

Reports from meetings

 

Architects Forum

 

The minutes of the AF meeting of 2 weeks ago were distributed to the MB.

A server with HyperNews, from SLAC,  will be installed by SPI and available to a restricted number of users. It is a pilot project in order provide a general bulletin board system to the experiments (that requested it at the AF).

 

Conclusions and follow-up from GDB

 

General points

 

A mail was distributed by K.Bos, MB and GDB members should send feedback.

 

Action: MB sends feedback on the GDB summary.


For convenience the mail is included below:


------- START EMAIL ------

[ . . . ]

-1- On the *GDB itself* in Phase II: the GDB discusses the problems and finds the solutions. The MB endorses the decision or resolves issues that are not decided at the GDB. Officially the GDB is a national representation but experts are always invited without previous notification. Meeting in April in Rome together with HEPIX. No GDB meeting in May and August. September meeting in Brookhaven.

-2- *VO Boxes were discussed again* as no clear decision could be reached at the Bologna meeting. It was agreed by the Tier-1s present (except NorduGrid) that VO boxes would be deployed at Tier-1 sites subject to completion and acceptance of the VO Box security and operational questionnaires (these should also be completed for VO specific services like PhEDEx). To provide a long-term strategy it was agreed that a task force be formed to look at commonality between VO Box services. To aid this group the experiments will provide sequence diagrams. KB is tasked with defining a procedure to take this work further forward The task force should meet as soon as possible and  the workshop in Mumbai should be used to discuss first results. Six months from now first evaluation completed and next steps decided.

-3- The October *Database Workshop was summarized* by Dirk Duellmann and a work plan presented. The plan contains very specific recommendations for hardware and software to be used for the service. There was no strong disagreement expressed with the plan so the 3D project will proceed as planned. Dirk will contact sites individually.

-4- Patricia presented the *new Gent4 VO* and asked the sites to support this VO. Generally agreed as valuable work for all LHC collaborations. Patricia was asked to follow the standard procedures for a new VO.

-5- Dave Kelsey presented *3 new documents for acceptance*: the Acceptable Use Policy Document, the Incident Response Document and the ??. The documents had been distributed well before the meeting and no comments were received. If no further comments are received within the next 10 days the Policy Documents are  agreed and accepted by all sites and MB endorsement will be asked.

 

 

Contributed by H.Marten:
- Grid acceptable use policy (V3.1, https://edms.cern.ch/document/428036 ready for approval)
- VO Security policy (V1.7, https://edms.cern.ch/document/573348 ready for approval)
- Incident Handling and Response Guide (V2.1, https://edms.cern.ch/document/428035 ready for approval)



-6- Roger Jones presented the results from the HEPIX/GDB *task force on storage*: studies of the computing models with detailed read and write rates and suggested hard- and software implementations. Archiving has not been addressed in full yet. Intermediate report this month. Next report at the HEPIX/GDB combined storage workshop in Rome after the April GDB.

-7- Nick Brook presented *what the experiments achieved* until today in sc3. General impression: software is stable but infrastructure is not. Lack of communication between sites and experiments on what is going on.

-8- Piotr gives an overview of all tests that are done on the *functioning on sites*. All services are published now and more information is added. Wish expressed to have a better understanding on how many cpu's are available for a VO.

-9- Alberto Aimar shows a *compilation* of the information given by the Tiet-1 sites of their plans for 2006. All data has been copied into a common template to ease the comparison. The template also contains the expected capacity and data throughput values. Some sites plan less than expected. All sites are requested to review their own data in the new template and correct or update when appropriate.

-10- Jamie presents *overall plans* for next year. In January repeat disk to disk throughput tests of SC3. Start in April and achieve nominal throughput values in July. According to what planning the Tier-1s have made available this cannot always be achieved. Criticism that there is too much focus on Tier0 data to tape at Tier1. Reading from tape for re-processing is not in the plans and neither the Tier1 - Tier1 or Tier1 Tier2 traffic has no milestone. Planning at the sites is difficult because of a lack of detailed knowledge what the experiments want to do. Need better communication on this between experiments and sites. TDRs and slides do not always show what is needed. We may need yet another group to address this? or can it be done between the VO and each side individually?

-11- The services at CERN are planned to continue over Christmas. LHCb wants to continue also, Alice and Atlas maybe and CMS no. KB will ask the sites for their services over Christmas.
 

[ . . . ]

 

So immediate actions from this meeting

- start a vobox task force with 1 person per expt and 1 techn. per T1.

 

- find someone to lead is and have a first 2-dy meeting in Amsterdam

 

- ask the experiments again to fill out the VOBox security questionnaire as well as the operations one

- last chance to react to the security policy documents

- ask Tier-1 to check their planning in the new format

- initiate discussion between experiments and one or two Tier-1 sites to discuss detailed flow diagrams to better understand what needs to be done

- ask Tier-1 sites if they plan to run over Christmas

 

[ . . . ]

------- END EMAIL ------

 

 

Proposed workshop at NIKHEF to define VO Box requirements


A workshop is proposed in order to define experiments’ needs and workflow.
The goal is to help Tier-1 sites to understand what is needed for the VO boxes and for other general services in the future.

 

C.Loomis will lead it and take responsibility that a formal statement of the requirements is clearly documented.

J.Templon will organize the meeting in NIKHEF.


Experiments and Tier-1 sites should attend. Not all Tier-1 sites must attend, but there must be enough (more than four) Tier-1 sites to discuss with the experiments.

January is considered a better time in order to have enough experience on the current grid services. Should not be later than that  because some requests may be needed for SC4, and therefore have to be implemented or modified before.

Due to the difficulty to have 2 consecutive days suitable for all 4 experiments, it was decided to have 2 x 1-day meetings. At each meeting there should be at least 4 Tier-1 sites studying the requirements of 2 experiments. Experiments that can participate to both days should do it: as that will help to identify and clarify commonalities. 

 

The MB decided to propose 19 Dec 2005 (LHCB + ATLAS or ALICE depending on availability) and  10 January 2006 (pre-GDB).


Action: L.Robertson will check that the dates (9 Dec and 10 Jan) for the workshop are suitable for C.Loomis.

Update: The dates are not suitable. The meeting(s) will have to be on the second half of January.

 

Getting up to speed on the Tier-1 data and work flows


Not all TDRs are easy to interpret. It is not easy to understand what it is really needed from the Tier-1 sites, both in terms of data and work flows.


It would be useful that the experiments provide “Tier-1 accessible” versions of the models described in the TDR. Should be pedagogical information targeted to the Tier-1 sites (the LHCB section with drawings and comments is a good example).  If the model includes also Tier-2 explanations that will be even better.


The experiments agreed to produce this material for the end of Dec 2005. Sites will then react and ask further questions in January.  The goal it to be able to wrap-up and converge during the SC4 Workshop on the Sunday before CHEP.

 

Action: By end 2005, the experiments provide a “Tier-1 accessible” description of their models, data and workflow.

 

AOB

 

No AOB.

 

Summary of the Pending Actions

 

New Actions:

 

  • L.Robertson and A.Aimar will prepare the High Level Milestones for formal approval by the MB.

 

  • MB sends feedback on the GDB summary.

 

  • L.Robertson will check that the dates agreed (9 Dec and 10 Jan) for the workshop are suitable for C.Loomis.
    Update: The dates are not suitable. The meeting(s) will have to be on the second half of January.

 

  •  By end 2005, the experiments provide a more “Tier-1 accessible” description of their models, data and workflow.

 

Pending Actions:

 

  • Tier-1 Sites (at GDB) should complete their plans following the guidelines of the presentation given at the GDB. (GDB presentation).

 

  • ATLAS will send a proposal about reporting on issues concerning non-EGEE resources to the MB.

 

  • CMS will send a similar statement to formally confirm the CMS opinion on OSG reporting.

 

  • Tier-1 Sites to set milestones on their known Tier-1-Tier-1 and Tier-1-Tier2 testing activities (IN2P3 stated specifically that they will do this in a few weeks).

 

  • All MB members - Preparation for the Mumbai workshop – details should be agreed following the 9 November GDB.

 

  • All MB members to provide more feedback on the High Level Milestone is needed.

 

 

The full Action List, current and past items, is available (see wiki page)