LCG Management Board

Date/Time

Tuesday 13 July 2010 16:00 - 17:00 - Phone Meeting  

Agenda

http://indico.cern.ch/conferenceDisplay.py?confId=81998

Members

http://cern.ch/LCG/mb.htm

 

(Version 3 -11.08.2010)

Participants

Jon Bakken, Ian Bird (chair), Kors Bos, Maurice Bouwhuis, David Britton, Tony Cass, Marco Cattaneo, Frédérique Chollet, Luca Dell’Agnello, Alberto Di Meglio, Dirk Duellmann, Michael Ernst, Sue Foffano, Qin Gang, Maria Girone, John Gordon, Denise Heagerty (notes), Andreas Heiss, Helge Meinhard, Gonzalo Merino, Harry Renshall, Yves Schutz, Jamie Shiers, Oxana Smirnova, Reda Tafirout, Romain Wartel

 

Invited

 

Action List

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

Mailing List Archive

https://groups.cern.ch/group/worldwide-lcg-management-board/default.aspx

Next Meeting

Tuesday 27 July 2010 16:00 - 17:00 - Phone Meeting

1.    Minutes and Matters arising (Minutes)

 

The minutes of the last meeting were approved.

 

2.    Action List Review (List of actions)

 

The action on John Gordon to provide dates for the storage accounting milestones is covered under point 4 below.

 

3.    LCG Operations Weekly Report

 

 

Summary of status and progress of the LCG Operations. This report covers the WLCG activities since the last MB meeting. All daily meetings summaries are always available here: https://twiki.cern.ch/twiki/bin/view/LCG/WLCGOperationsMeetings.

 

Jamie Shiers presented the slides attached to the agenda page. The following comments were noted:

Slide 3: concerning the ATLAS-BNL plot in red, Maria Girone agreed to follow up.

Slide 12: a10Gbps switch (not a 10Gbps link) became unavailable.

Slide 17: to complement the points listed on the Collaboration Workshop draft actions slide, the following comments were made:

  • possible rotation of the chair of the Tier-1 coordination meeting amongst the Tier-1 sites
  • more consistent use of microphones in the bi-weekly service coordination meeting (to help remote participants)
  • less in depth problem solving of issues specific to the Tier-0 during the daily and bi-weekly meetings

 

 

4.    Proposed milestones for storage accounting

 

 

John Gordon proposed the following milestones for storage accounting:

  • Early Sep: verified reports available for Tier-1 sites
  • End of Oct: verified reports available for Tier-2 sites

These milestones will be tracked by the MB. Action: John Gordon to report on storage accounting milestone status in early Sep and End of Oct 2010.

 

As ALICE migrate to native xrootd, their storage information is not available in the BDII for generating the storage accounting reports. Action: John Gordon will follow up on provision of the xrootd information needed for the storage accounting reports by the next GDB (8 Sep 2010).

 

 

5.    Status of 2010 pedge installations

 

The following table gives the status of pledged resources at the Tier-0 and Tier-1 sites.

 

Site

CPU

Disk

Tape

CERN

100% (but 15% in use for stress tests)

100%

100%

TRIUMF

100%

100%

100%

CC-IN2P3

100%

100% by 1 Sep 2010 (but allocated on demand)

100% (cartridges purchased as needed)

GridKA

100%

100%

100%

CNAF

100% by 1 Sep 2010

100% by 1 Sep 2010

100%

NL-Tier-1

100%

100% (but not all allocated yet)

100% (but slowly increased)

NDGF

100%

100%

100%

PIC

100%

100%

100%

ASGC

100% by 14 July 2010

100% by end of Oct 2010

100% by end of Oct 2010

RAL

100%

100%

100% (but just in time delivery)

BNL

100%

100% (plus an extra 20%)

100%

FNAL

100%

100%

100%

 

Kors Boss clarified that as ASGC is currently in validation state, their plans to install full capacity by end of Oct 2010 is acceptable for ATLAS.

 

Sue Foffano reminded the sites that pledge input for 2011 will be needed by the end of Sep (for the RRB meeting in Oct). She will send email reminders.

 

 

1.    Quarterly Reports (2010Q2)

 

 

The Experiments were asked to report on issues during 2010Q2:

 

ALICE: no particular issues to raise.

ATLAS: no particular issues to raise.

CMS: not present.

LHCb: investigating queue lengths as jobs are taking longer than expected.

 

The missing input for the 2010Q2 quarterly report is requested by end of this week.

 

 

2.    Follow-ups from workshops

 

 

Ian Bird presented a slide of actions from recent workshops (DAaM jamboree, multi-core & virtualisation workshop, WLCG collaboration workshop). There were no additions to the list of demonstrators. Work on the SRM speed-up is straight forward and is already underway. The demonstrators will be reviewed at the Oct GDB. They will need to provide their status and milestones of what they expect to achieve by the end of the year.

 

Ian Bird asked for a clear statement from the Experiments, in the next couple of weeks, on the following 3 points:

  1. Xrootd is a strategic direction for data access for all experiments?
  2. Whole-node reservation is a clear requirement? For all experiments?
  3. CERNVM and/or CVMFS is a strategic direction for all experiments?

 

Action: Experiment Computing Coordinators to provide a statement on the above 3 points

 

Ian Bird proposed to set up a small architectural team to outline an overall strategy for the following:

 

1. Define subset of SRM needed for archives

2. What is needed for file transfers? FTS/http(s) ???

3. How to determine data hotspots?

4. How should catalogues evolve?

5. Better monitoring + better use of monitoring resources

6. What is WLCG strategy for use of virtualisation?

7. End-end changes to support multicore, etc.

8. Squid/Squid+Frontier service?

& etc.

 

The use by ALICE of xrootd (instead of FTS) for file transfers is creating issues for sites since it reduces their ability to monitor and this traffic. Ian Bird will agree with Yves Schutz a set of questions to be answered by ALICE to address this concern at a future MB meeting.

 

Action: Ian Bird to create a small architectural team

Action: Ian Bird to define xrootd related questions to be answered by ALICE at a future MB meeting.

 

 

 

3.    AOB

 

 

3.1     Security incident

An ssh security incident is currently being investigated in the WLCG community. All sites have been asked to take action and check for patterns that have been sent to them. Romain Wartel asked sites to make sure that their security representatives are following up on these requests. With the move to EGI, it is important to ensure that information has been received correctly by all sites.

3.2     Next meeting

The next meeting will be on Tuesday 27 July 2010 (phone meeting).

 

 

4.    Summary of New Actions

 

 

  1. John Gordon to report on storage accounting milestone status in early Sep and End of Oct 2010.
  2. John Gordon to follow up on provision of the xrootd information needed for the storage accounting reports by the next GDB (8 Sep 2010).
  3. Experiment Computing Coordinators to provide a statement on the following 3 points:

a)      Xrootd is a strategic direction for data access for all experiments?

b)      Whole-node reservation is a clear requirement? For all experiments?

c)      CERNVM and/or CVMFS is a strategic direction for all experiments?

  1. Ian Bird to create a small architectural team.
  2. Ian Bird to define xrootd related questions to be answered by ALICE at a future MB meeting.