Dear All,
Please find attached the last three weekly GridPP Project Management
Board Meeting minutes. The latest minutes can be found each week in:
http://ppewww.ph.gla.ac.uk/~doyle/GridPP/PMB/minutes/this-week.txt
(as well as being listed with other minutes e.g. 011029.txt, 011109.txt,
011112.txt).
Cheers, Tony
PS Apologies for the delay in sending the earliest of these out.
Feedback on the format of the collaboration meetings welcome, as noted
in the second of these minutes (1-day meeting at RAL: 011109.txt).
________________________________________________________________________
Tony Doyle, GridPP Project Leader Telephone: +44-141-330 5899
Rm 482, Kelvin Building Telefax: +44-141-330 5881
Dept of Physics and Astronomy EMail: [log in to unmask]
University of Glasgow Web: http://ppewww.ph.gla.ac.uk/~doyle
G12 8QQ, UK Video - IP: 194.36.1.234
________________________________________________________________________
GridPP PMB Minutes 8 - 29th October 2001
=========================================
Present: Roger Barlow (RB), Dave Britton (DB), Nick Brook(NB),
Tony Doyle(TD), John Gordon(JCG), Steve Lloyd(SLL),
Robin Middleton(RPM), Les Robertson(LR)
Review of Actions
-----------------
1.11 ongoing
3.2 ongoing; recieved post estimates from all expts by today's deadline -
see tendering process below
4.7 Have the info now and will get OK from KJP as soon as possible.
5.7 ongoing
6.1 done (by SLL)
6.2 ongoing
6.3 ongoing (all comments to the PMB list)
7.1 done
7.2 done
Tendering Process Discussion
----------------------------
Requests for new posts...
Atlas/LHCb 4 posts (2 on GANGA, 1 on installations, 1 each for MC
production - the first 3 in common with LHCb)
BaBar - 5 areas in job submission, metadata, certs/auth, obj, data
dist & replication
CMS - 5 posts (2 in devs for regional centres, 2 posts on MC, 1 on
object replication
CDF - 3 posts (2 for adaptation of SAM to CDF, 1 for dev of GUI for
data catalogue browsing system)
D0 - 4 posts (3 for generalisation of SAM for other users, 1 on MC
production)
LHCb - see ATLAS submission
other - 2 posts for data discovery, etc
UKQCD - 2 posts (job submission, data replication, etc...)
Need more clarification on BaBar and CDF/D0
The TAG have a working document, but not yet ready for release. (recent
new input from JCG)...thus question on whether to delay the timescales.
However, since end-points are fixed we can't really play with the
schedule. The TAG document is not too usuful after the tendering process.
Thus final proposal is that the TAG document is on open release before 9th
November 2001 (or else there will be no document publicly released) and
semi-decouple from the tender document which should still go out on 2nd
November. Originally the tender document was to define quite tightly what
would be requested, but this will not be possible in such detail if the
TAG document is not fully available.
Possibility of some common work between experiments has been mooted,
though at this late stage in the process it may be hard to achieve.
BaBar was defined as leading the excercise, but this was not deemed
acceptable by the others involved. There still remains the common
requirements in terms of job submission, data management and getting
outputs back.
The point was made that experiments should beallocated posts and then
asked to advise where they should go, rather than if the guidelines are
generic and the institutes asked to bid directly (which would result in a
wealth of applications from institutes to increase chances in the final
outcome).
Thus, 6 LHC, 6 non-LHC and 3 generic is still the general proposal (the
existing 5 posts in WP8 of DataGrid would also be in the LHC area).
A working allocation was finalised by Thursday that week.
ACTION: TD to circulate first draft of post allocation for comment and
finalisation by Thursday.
Point made that 1 extra for Tier-1 is not enough ... the current 9 + 1
will not be enough to support the expanded role of a Tier-1 centre with
international obligations. John has submitted an overall envelope of 14
for the Tier-1, but if the allocation is more likely to be 11 then CLRC
would have to say what could be done for that amount of effort.
ACTION: NB - to clarify the overlaps between experiments
CERN Procurement (see proposal)
-------------------------------
Les assumes that CERN buys the equipment and then charges to PPARC, but
this still has to be confirmed by NIG. However, CERN needs urgently to
start the long tendering process. The proposal includes infrastructural
costs (racking, network, cooling, installation, etc.). Disk prices may be
a bit low, but the estimate may need revision taking into account raw .vs.
useable space.
DECISION: CERN procurement approved by the PMB.
Next Collab. Mtgs
-----------------
Concerns in fitting in with other experiments, GGF and in particular the
external review of DataGrid by the EU. Priority would be 1) 14/15th Feb
2002, 2) 21st/22nd Feb 2001 and last 28th Feb 2002 /1st Mar 2002
ACTION: TD to consult Andy Parker on dates and then get back to the PMB.
The meeting next week is confirmed...the website is available for
bookings...people will be in various B&Bs around Edinburgh. VC facilities
are subject to confirmation. PMB will be prior to dinner in NeSC.
ACTION: TD to ask Steve Playfer to circulate known participants list via
UKHEPGRID list.
Future Meetings
---------------
PMB on 9th November at RAL will start at 10am (coffee at 9:30am) - JCG will
book and organise ISDN VC for Les and Roger.
GridPP PMB Minutes 9 - 9th November 2001
=========================================
(1-day meeting at RAL)
Present: Roger Barlow (RB), Dave Britton (DB), Nick Brook(NB), Pete Clarke(PC),
Tony Doyle(TD), Neil Geddes(NIG), John Gordon(JCG), Steve Lloyd(SLL),
Robin Middleton(RPM), Les Robertson(LR)
(Roger and Les via VC)
A short (1 hour) meeting of the PMB at the Edinburgh GridPP Collaboration
Meeting on 5th November was used to gather requirements for this 1-day
meeting.
Review of Actions
-----------------
1.11 PC & DB: Project management tools still under investigation.
MS project was used for examples of experiment timescales (presented by
PC at the Collab. Mtg.)
3.8 NIG: Oversight board meeting confirmed 14th December
Review - resource choices, deliverables, status of project
NEW ACTION: Definition of high level deliverables: annual & quarterly
(TD, JCG & PC)
4.1 TD: PMB web pages still to be done. Subsequently completed by SLL.
see http://www.gridpp.ac.uk/pmb/ and accessible via "About GridPP" link
4.7 RPM: to provide cost of LTA at CERN
Done
COLA - £5100 p.a.
London weighting - £1760 (subject to clarification)
+ accommodation (at reasonable cost)
5.7 PC: List of names of TB: ongoing - subject to discussion during meeting
6.2 JCG: H/W specs to EB will be done next week
8.1 NB: Clarify experiment overlap
Done.
CMS/BaBar objectivity identified as a possible common project.
8.2 TD/SLL: date & venues for future collab. meeting
Done.
Feb 14/15 - Cambridge (Andy Parker)
May - Manchester (RB & Andrew McNab)
November - Imperial (DB and Dave Colling)
Feedback given to organisers of Edinburgh meeting
Annual £30k budget for Collab. Meetings (3 *10k)
Proposal that future collab. meeting agenda be
organised with 2 sessions for Experiments overviews (EB)
& 2 sessions for Technical Status (TB).
Request for input to TD from all members of GridPP on required format for
future collaboration meetings (with increased emphasis on technical
developments).
Experiments Board - NB
----------------------
NB presented his & RB's view of the role of the EB and relationship with
the TB now we are moving beyond tendering stage.
feedback from experimental project
monitoring of expts. (TB X-members important role here)
dissemination between expts.:
- inside GridPP
- outside GridPP
meet every 3 months (i.e. same as "reporting" schedule )
Relationship with TB
Feedback & prior discussion of:
a. middleware issues
b. development at the "Tier" centres
Detailed assessment of expts. technical input
Following discussion of this proposal, PC presented his view on the role
of TB.
Technical Board - PC
--------------------
The TB will have responsibility for monitoring of GridPP 3-monthly
deliverables, which includes the experiment deliverables, with no formal
role for the EB in this procedure.
Following discussion, the PMB approved this general approach.
The role of the EB was discussed further, without reaching specific
conclusions. NB will consult with the EB on its role, matched to the
EB role specification, assuming that the TB is responsible for 3-monthly
monitoring.
The make-up of the TB was discussed and names were suggested for membership.
The TB will organise technical meetings e.g. EDG testbed1 roll out, Data
management "how to", etc.
PC will start process of iteration with expts to refine milestones &
deliverables.
The issue of how best to determine Tier 1/A milestones was discussed.
Andrew Sansum & JCG will produce a proposal. This led to a discussion of
displacement, in particularly at CLRC. NIG noted that displacement at RAL
had to occur this year in order to avoid cuts in PPD & ITD. The only
people to be displaced this financial year will be from RAL.
ACTION - the deliverables & objectives of GridPP buying effort from PPD &
ITD should be defined (i) for this year & (ii) for the duration of the
project. TD to contact Ken Peach - with a deadline of the end of
November, in order to provide input to the GridPP Oversight Committee
(see below).
Resources - DB
--------------
DB summarised the financial commitment and presented the spreadsheet tools
developed to monitor the project's financial commitments. There was
agreement to provide education & training on issues with direct relevance
to Grid activities. For current CNAP courses (e.g. OO, C++) PPARC
e-science will provide funding. Steve Fisher to be encouraged to go ahead
and organise courses as usual.
It was noted that it might be possible, in some circumstances, to use Tony
Hey's core e-science funds to finance travel for GGF.
The deadline for the BaBar common fund rebate (offset against the Tier A
centre) is in January. RB pointed out in order to qualify for this BaBar
UK must make a commitment to have access/use of 2/3 of the GridPP h/w
resources. Discussion followed on what this actually meant in reality.
ACTION: RB to submit the BaBar UK Tier A proposal to PMB for discussion
before final submission to BaBar.
DB will build into his financial tools, the numbers and integrity
checks with extraction & reporting mechanisms. Following on from the BaBar
discussion it was noted it will be important to monitor use of resources -
JCG notes that a CNAP-like mechanism at Tier 1/A is all that could be
currently envisaged. This required further consideration.
CERN - LR
---------
CERN is in the process of setting up the 3 LHC computing committees
(Overview, PEB, SC2). The expts will nominate their representatives to
Hans Hoffmann by Nov 12th. It was hoped there would be an initial meeting
of the SC2 before Christmas and a workshop is planned to be organised for
Feb 2002. The s/w co-ordinators of the LHC expts have already requested
that some of the Technical assessment groups be set up before the
convening of the SC2 eg assess persistency solutions. The PEB will meet in
the next 10 days. GridPP should have representatives on all 3 committees.
Preliminary names, discussed at the meeting were:
Project Overview Board (Chair: Hans Hoffmann) - Neil Geddes
Project Execution Board (Chair: Les Robertson) - Tony Doyle
Software and Computing Committee (Chair: Matthias Kasemann)- Pete Clarke
On CERN recruitment, 29 applications were judged to be above threshold.
They will be interviewed in the UK on 3rd/4th December. Other interviews
will take place at CERN on 6th December.
Tier 1/A & Testbed - JCG
------------------------
JCG has produced a Tier 1/A requirements document for the GridPP tendering
procedure. The tender for the h/w of the Tier centre has gone into the EU
journal. It is foreseen that the evaluation will begin the end of
December/early January with delivery the end of January/early February. He
will circulate the specification document to the EB next week. The interim
disk purchase will be requested.
For the testbed roll-out, RAL seems to be slipping down the list of
priority sites in EDG. The testbed h/w is ready to go and will gain
experience of Globus 2 (outside of EDG) by installing the toolkit
independently.
DataGrid - RPM
--------------
Busy preparing for EU review of DataGrid. RPM will contribute input on
EU-funded staff at RAL to the Q4 and annual reports. Due to the late turn
on of GridPP funding, problems with PPARC contributions to DataGrid. WP3
contrib. is now OK. WP5 is 1-2 quarters behind schedule. RPM outlined the
WP5 recovery plan: (i) immediate/short term: use of agency staff to get
things going (ii) medium term: offer 6-12 month contracts. (1 verbal
acceptance at the moment). Make use of other recruitment in CLRC to work
in WP5 area. (iii) general recruitment: advert for 2 EDG funded posts
(WP3 & WP5) and 2 EDG unfunded posts (WP5 & WP6) is running currently and
closes next week. So far 25 applicants for posts. Highest priority to
recruit WP5 posts. Aim is now to have 3 staff years over 2 years.
NIG stressed it was essential that the UK EDG was seen to be putting its
house in order and being pro-active in addressing the problems of WP5.
In the worse case scenario PPARC could be thrown out of EDG if it's felt
that it's not meeting its commitments. It was suggested the current CLRC
WP8 effort be temporarily re-directed to work on WP5. NB expressed concern
on this development and thought it would have consequences on the roll-out
of the testbed at RAL and its subsequent evaluation by UK LHC
collaborations. It was noted that CMS UK had already moved their UK
priority site from RAL to IC over concerns over lack of expt support.
There were ~8 "unfunded" (i.e. GridPP funded) EDG posts in position now.
Concern was raised about the lack of a UK political presence in WP6,
since Dave Newbold stood down during the summer. A suitable candidate to
take over from Dave needs to be found.
AoB
---
PC raised the issue of preparation for the Oversight committee. He
suggested a dedicated PMB meeting to plan the presentation. He raised the
question of how we address the cut-backs to the proposed programme
due to the reduced funding.
Meeting closed at ~16:30
The next GridPP PMB 1-day meeting is planned for Wednesday 16th January at
Imperial College. This should include time for further discussion of the
EB and TB roles and their relation to the PMB.
GridPP PMB Minutes 10 - 12th November 2001
===========================================
(1/2 hour follow-up to 1-day meeting via VC)
Present: Dave Britton (DB), Nick Brook(NB), Tony Doyle (TD),
John Gordon (JCG), Robin Middleton (RPM).
A short discussion on preparations for the reports/presentations to
the GridPP Oversight Committee:
It was agreed that three members of the PMB would present information:
Tony Doyle - overview;
Pete Clarke - technical progress report;
Dave Britton - resource management progress report.
In addition, Steve Lloyd will be present to report back to the
Collaboration Board.
The presentations would discuss the current status of the project,
including the CERN component, and outcomes from the PRSC, the week before.
A short discussion followed on the presentation of the balance between an
experiment-led project, key middleware deliverables and establishment of
an integrated UK testbed.
Details of the requirements are currently being discussed with the
Oversight Committee members and will be fed back to the PMB shortly via
Guy Rickett. Documents would be prepared according to these requirements.
Note added: the role of the Project Oversight Committee, including
proposed reporting periods at 6-monthly intervals, was sent by PPARC
to the PMB later that week (14/11/01).
ACTION: ALL to provide input for discussion of format of documents
required by the oversight committee.
Actions (12/11/01)
------------------
1.11 PC & DB: Project management tools.
First versions required prior to the Oversight Committee Mtg.
5.7 PC: List of names of TB: ongoing - subject to discussion of TB role.
6.2 JCG: H/W specs to EB will be done next week
9.1 TD/JCG/PC: Definition of high level deliverables: annual & quarterly
9.2 TD: to contact Ken Peach on RAL deliverables & objectives for GridPP
existing effort from PPD & ITD (i) for this year & (ii) for longer term.
9.3 RB: to submit the BaBar UK Tier A proposal to PMB for discussion
before final submission to BaBar.
10.1 ALL: to provide input for discussion of format of documents
required by the oversight committee.
|
