Dear All,
Please find attached the F2F and weekly GridPP Project Management
Board Meeting minutes. The latest minutes can be found each week in:
http://www.gridpp.ac.uk/php/pmb/minutes.php?latest
as well as being listed with other minutes at:
http://www.gridpp.ac.uk/php/pmb/minutes.php
Cheers, Tony
________________________________________________________________________
Prof. A T Doyle, FInstP FRSE GridPP Project Leader
Rm 478, Kelvin Building Telephone: +44-141-330 5899
Dept of Physics and Astronomy Telefax: +44-141-330 5881
University of Glasgow EMail: [log in to unmask]
G12 8QQ, UK Web: http://ppewww.ph.gla.ac.uk/~doyle
________________________________________________________________________
GridPP PMB Minutes 243 (F2F) - 18th January 2007
================================================
Present: Tony Doyle, Sarah Pearce, Roger Jones (Minutes), Stephen Burke,
David Britton (Chair), Dave Newbold, Steve Lloyd, John Gordon, Andrew Sansum,
Neil Geddes, David Kelsey
Apologies: Robin Middleton, Jeremy Coles, Peter Clarke, Glenn Patrick,
Deborah Miller, Tony Cass
1. Status of Quarterly Reporting
=================================
It was noted that preparations for the OC Quarterly Reporting are
required of Dave Britton by Wednesday at the latest. It was reported that
applications reporting exists for ATLAS, CMS, QCDGrid and GANGA. There is
an understood delay for LHCb. The others will need chasing. The status of
the middleware reporting is not known, other than that the security report
has been sent. The Tier 1 and 2 reports are still needed. There may be
problems in WMS; Dave Colling was going to talk to EGEE colleagues to
redefine the project scope and plan. The interoperability also looked
problematic at first glance, but this has probably been resolved in
discussions this morning.
2. Previous OC Minutes
=======================
a) GridPP Roadshow for 2007
----------------------------
Regarding a possible GridPP road show in 2007, it was noted that this
could be equated to the site readiness review, communicating with system
administrators. The user base is being addressed by tutorials and by
presentations at the IoP meeting. It was agreed that DK will e-mail
Darren Green for a contact e-mail with goal of future talks with industry.
We are also in contact with Price Waterhouse Cooper.
b) User Board in GridPP3
-------------------------
There is a request for clarification of the role of the User Board in
GridPP3. DB reminded us that the UB and Deployment Board will have more
cross-membership, and the DB will be more user-focussed. The proposal is
the the Chair of the UB be appointed for the duration of the project, and
be bought-out at 0.25 FTE. What is the mechanism for appointment? DB
thinks it should be a two-way role, and so needs the confidence of the
management and the users. There is a proposal that there could be an
election (which may not satisfy the management), a search
committee/consultation (which might satisfy both) or a PMB discussion
(which might not satisfy the user community). The feeling was there should
be a search committee. This is not possible until after the funding
position is known, so this will probably happen in April.
c) GridPP Usage Statistics
---------------------------
There were also questions about usage statistics. JG presented some of
the resource usage statistics from JC. The reported utilisation (even
after efficiencies are taken into account) are not at the 85% target
level. The disk usage is low. In particular, Manchester declare a huge
amount, but not much is used, and the accessibility is not clear. This
needs follow-up. The plots are not well understood and are not for general
dissemination. JC also reported on the monitoring system. There is
still a need to remove any manual steps. The availability statistics look
reasonable (although require careful reading). Sites request that
experiment-specific tests are integrated with the CERN tests, as they do
not want a plethora of tests running.
The feeling was that the most important item is the outturn. SL is
quite well advanced with this, but needs to understand the Tier-1 numbers.
There was discussion if the questions from the OSC about the ability to
track usage (and users) in our response. TD showed some slides on the
number of active users (>10 jobs run) and the number of registered users.
The fraction of active users is rising but still only 11%. There was a
long discussion on relevant metrics and appropriate targets. The current
situation is still dominated by production rather than user analysis so
the number of active users is modest.
3. Dissemination Report
========================
SP reported on upcoming events:
- IoP HEP group in Guilford - we hope to take the 3D RTM and posters
- LondonGrid meeting - this is to encourage other communities.
- QMUL opening by Alan Sugar in May.
- OGF and EGEE meetings in May. We will have bloggers and take the usual
RTM/posters. Abstracts have been requested, but abstracts from sites on
their experiences are also invited. Ian Bird would like to know of such
abstracts.
- CHEP 2nd September
- BA Festival starts 12th. Neasan O'Neill will set things up here after
the All Hands stand is set up.
- All Hands 10th-13th September. This will be the first for the STFC;
there were three stands in its remit last year, and this will probably
reduce to two. However, the stand space has already been allocated, so
be may arrange a double space.
- We need to be sure that LHC@Home has lots of jobs ready to run, so
people get something when they subscribe after seeing our publicity.
There is hope to have other programmes to run on the system.
- The first International Science Grid This Week editorial board will be
next week, with SP as a member. Three UK articles are currently planned.
- We have 3 industry projects ongoing. Cambridge Ontology are just
starting. Total are also just about the enable the VO.
- QMUL are starting a new project with an econophysics spin-out.
- There is the question of a GridPP2 paper. There was little enthusiasm,
but there is a case for several small papers. There may also be an EGEE
topical journal edition.
4. Dissemination Funding for GridPP2+
======================================
There is a question as to what we do with dissemination funding in the
GridPP2+ era. SL has to project the existing money to see how long it
will last. A request to PPARC is needed to allow the reallocation of other
funds if necessary.
5. LCG Milestones
==================
The current milestones are based on the old LCG plan. Jamie Shiers is
taking the experiment plans and trying to reverse-engineer the
requirements. It is hoped that the real experiment use will allow many
existing metrics to be met. It is clear the experiment models require
various simultaneous activities (T0-T1, T1-t1 & T1-T2 transfers at various
rates). It is hoped that the experiment tests will cover this. It is not
clear that Jamie has decided which scaling of the full rates is needed in
07Q1. Discussion of the OPN topology revealed that the computing models
and the megatable have still not been properly absorbed/aligned with
peoples thinking.
The data management/SRM group will be looking at internal access patterns.
There is a need for a sharing of data between ATLAS and CMS on castor
access, and more discussion of the internal flows in the Tier-1. We should
arrange a dedicated half-day meeting.
6. UK Site Readiness Review
============================
A Site Readiness Review of all GridPP sites is proposed. Everyone on the
PMB will need to be involved. Each site visit will need a representative
of the management, of the experiment and of the technical side. There was
prolonged discussion of the teams. There was also a discussion of how a
separate team can allow a fair judgement of all sites. The opinion was
that the review should be on matters of fact, and so should not be
contentious if not always comfortable. It was suggested that the T2
co-ordinators could also be involved in the review of other sites.
Northgrid: TD, GP, JC, PG
London: RM, DN, SB, GS
ScotGrid: DK, RJ, JG, AF
SouthGrid: NG, SL, AS, OA
DB will co-ordinate the whole activity. The team will meet with the
Tier-2 Co-ordinator, the local sys. admin and the local PI. The review
should cover fixed topics and metrics. TD has a prototype questionnaire
that can be developed. It covers hardware, software, operating systems,
users for the resource, configuration management.
The plan would be to have the reviews completed by the end of April.
7. EGI and EGEE Future
=======================
RM commenced by summarising the EGEE3 developments and GridPP's role
within it. The call is expected to open in May and end in the autumn.
The idea of centres of excellence is on the agenda, and there will be
pressure not just to spread the cake geographically.
EGEE activity leaders have produced proposals for each area, optimising
the use of resources and modifying structures where necessary to better
achieve goals. Some new ideas around automation to reduce effort in
operating the Grid and in instrumenting the operational infrastructure
producing data for use by computer scientists are under discussion.
Significant changes in interaction with industry are also under review.
Middleware maintenance is expected to be a major priority above middleware
development.
The maximum EU funding moves from 50% to 75% for most partners in EGEE,
but we are likely to see the same funding in Universities and labs. If
adopted at "face value", just over half of the current EGEE2 effort could
be afforded! The EGEE PMB therefore committed to a project of the EGEE-II
scale, so the institutes will put back roughly 25% in matching funding.
The end result will be very similar to the present situation. Partners
should declare if they are in two different bids competing against the
same pot of money. It was noted that Joint Research Units (JRU) are
required and must be in place before the project.
There is a shift from GridPP to NGS. The JRU has to have a longer
existence than EGEE3; the NGS is best placed to undertake this, though
most likely in tandem with GridPP. Does this leave GridPP visible enough?
There are risks, but also opportunity. The JRU will be lead by CCLRC/STFC,
and the money will be distributed by contracts. There will be a precursor
project for the EGI design.
There was then a debate about the GridPP position; do we want to bid for
the Tier 2 co-ordinators or for shadows? We presumably still need an
operations co-ordinator. This is agreed. We might part-fund the regional
co-ordination, but more is needed to cover the other NGS sites. History
means that the existing T2C hosts are partners. On the middleware side,
security needs to be bundled in. How does OMII-UK and OMII-Europe fit into
this? The UKI view is that there should be no middleware development in
EGEE3, but there should be middleware support from somewhere. The priority
is to protect the operations over other tasks.
8. Hardware Purchase Plans for 2007
====================================
It was noted that GP was absent, and AS will discuss this at the T1B
tomorrow. An agreed timeline is need. AS would also be interested in a
framework agreement when substantial funds are obtained. However, it has
timed-out for 2007. The feeling of the European Tier 1s are that you
should tender each time. There will be a decision tomorrow, but we should
give full consideration to risk management/minimisation. The T1B must come
to an agreed position, but may not be in possession of the full facts.
It was noted that we are undersupplied in November 2007, and this could be
bought in from 2008 or we could just work for the June 2008 target. Should
we continue with an existing supplier if we buy a smaller purchase? There
is also pressure to buy early to reduce risk, but there will be a hit in
cost. One possibility is to split the purchase and stagger so we always
have some insurance.
9. R-GMA Situation
===================
It was noted that the R-GMA support in GridPP2+ was reduced from 3 to
1.5FTEs. The EGEE contract situation had not been considered, and so a
recovery was allowed. A proposed GridPP+ fix using 31.6k of saving was put
forward and we are allowed to do this. DB proposes we find that money in
GridPP2 to allow us to deliver in GridPP2+ to within 12 staff-months of
the agreement. This should allow an orderly completion of the major
deliverables. Once the future of GridPP3 is known, the R-GMA plans will
need to be adjusted. This was agreed.
10. AOB
========
None.
GridPP PMB Minutes 244 - 29th January 2007
===========================================
Present: Tony Doyle, Sarah Pearce, Roger Jones, Stephen Burke, David Britton,
David Kelsey, Dave Newbold, Steve Lloyd, Tony Cass, Robin Middleton,
John Gordon, Jeremy Coles, Peter Clarke, Glenn Patrick, Andrew Sansum,
Suzanne Scott (Minutes)
Apologies: Neil Geddes
It was noted that Yingqin Zheng was not present at today's meeting.
1. Review of OC Documents
==========================
TD commenced by suggesting the meeting go through each of the draft
documents in turn, each author to provide a 5-minute review. This will
allow a check on any overlaps and omissions.
a) LCG Status Report
---------------------
TC reported that there had been steady progress towards delivering a
stable production service, but that performance targets had not been met.
It was noted that the gLite Resource Broker was not quite stable enough to
use yet, and there was a shortfall in hardware funding. There were
improvements with CASTOR and it was now deployed at RAL. There had been
tape acquisition at CERN. Steady progress was made, but performance
levels need to be met round the clock.
TD noted several points that could usefully be amended:
- at the start, the section on developing site reliability, it needed to
be made explicit that this refers to Tier-1
- planning for 2007 needs to be referenced, along with a possible mention
of the WLCG workshop that GridPP had fully contributed to
- the comment that 'SC4 performance targets were not fully met' (on p2)
should be explained or even omitted, with the sentence beginning:
'Transfers from CERN ...'
- regarding the EGEE Resource Broker, in the context of EGEE this is
understood, but there is implied criticism here, that the Resource
Broker is not meeting requirements. It was known that the code is not
stable and there are delays in testing, but are these recognised in the
EGEE report?
- On the Tier-1 side, a question may be asked about the budget being
sufficient - delays in LHC timescales affect this, so a statement needs
to be made about whether the budget can meet 2007 plans and into 2008
- a 'Summary' heading is required at the beginning of the last
(concluding) paragraph, with 'Repeating the introductory sentence'
omitted
- regarding overall planning, TC will investigate and add something, also
noting that the extension of testing to the Tier-2s is being considered
for 2007
b) EGEE Report
---------------
RM reported on the EGI and Design Project. Regarding EGEE-III, the Joint
Research Units (JRU) are the stepping stone to forming NGIs. It was noted
that technical discussions were underway. RM noted an EGEE-I and II
retrospective including an EU review. EGEE visibility was raised at the
UK e-Science AHM and through the User Forum in Manchester. RM discussed
Middlware and the R-GMA situation.
TD noted that one point to add regarding EGEE-III and EGI, was GridPP3's
position on this as stated in RJ's Minutes from the F2F meeting - his
statement captured the discussion, with an emphasis on operations. TD
suggested that the wording could be extraced from there. TD also noted
the omission of 'probably' at the top of p3. It was agreed that the F2F
Minutes would be circulated [this was done following the meeting].
It was noted that JG had a paragraph to add to the EGEE Report on
Operations. Similarly, paragraphs on Security and Vulnerability were
required. TD noted that the old structure should be preserved and used in
the MSN Report. Responsibilities could be mentioned here but described
more fully within the MSN Report.
c) Deployment Report
---------------------
It was noted that DK had circulated an email outline providing an overview
of the year's activities. Issues comprised current metrics, deployment
meetings, future plans, the Site Readiness Review and keeping to MoU. TD
noted that the plan was good. DK noted that he now had all the required
information, he just needed to write it up.
d) MSN Report
--------------
RM reported that he was taking input from the last Quarter, and needs to
speak to TD/DB regarding the outcome of GridPP2+. He also needs to review
feedback and review the middleware development activities. It was noted
that the Report was difficult to write in relation to GridPP3 as areas are
likely to be curtailed. RM will speak to TD sometime on Tuesday and
circulate a draft Report by Wednesday.
e) Applications Report
-----------------------
It was noted that RJ was not yet present. DB reported that he had done
some follow-up work on applications, and everything on the LHC experiment
seems under control. There is however a problem on the BaBar side where
there is a need to place deliverables on the project map but there are
issues to be resolved. Portal and UKQCD is OK but there is a lack of
portal applications following GridPP2. RJ had sent a draft summary of the
report to DB but his schedule for completion is unknown. DB noted that a
statement needs to be made regarding applications being cut, however an
approach needs to be made to PPARC regarding this. It was agreed that DN
will circulate a draft statement for discussion/approval.
f) User Board Report
---------------------
GP noted that resource requirements had departed from the quarterly
assessment. He took an overview of 2007 in his Report. It was noted that
BaBar could not meet their requirements; Tier-1 had sufficient CPU
capacity and Tier-2 resources look adequate. GP reported on resource
allocation and the Data Challenges. The Report also focussed on CPU
efficiency, Tier-1 issues, migration to CASTOR, Grid-Only Access, and the
future of the User Board.
TD noted that regarding resource requirements at the Tier-2, it might be
useful to refer to GOC accounting and perhaps rephrase the last sentence
regarding Tier-1 statistics. For resource allocation, insert 'March 2006'
procurement (instead of 'various') so that this tallies with the Report by
AS.
Regarding the section of Experiment Data Challenges, this is also
referenced in the Tier-2 report. The statement beginning: 'Since early
December .... Pentium 3 machines' should be omitted.
Regarding the CPU efficiency plot, AS has a high-level summary. TD
suggested omitting the plot or commenting on context from individual users
on the overall efficiency of the ~91% and falling to ~77%. There was a
discussion as to whether this plot and the other one should appear in AS's
Report, and be discussed there. AS is taking an overview of Tier-1,
whereas GP is taking an overview of the experiments. It was agreed to
refer to the Tier-1 report and comment here on the complexities of these
CPU efficiency figures, and that the efficiencies will never reach 100%
Regarding the 'Future Arrangements' for the User Board, it was noted that
this proposal was subsequently approved by the PMB, and the wording should
be amended to reflect this. GP was asked to review and standardise fonts.
g) Tier-1/A Report
-------------------
AS provided a summary of deliverables, then usage and outturn, and disk
resource problems. It was asked if this overlapped with experiment
requirements, but both were fine and consistent. It was noted that 1.5
million was available in GridPP2 to tender for disk and CPU - iteration
can provide the fraction, but this could be in the Resource Report by DB.
It was noted that deployment and infrastructure are going well, the CASTOR
deployment is challenging but the target rates are being achieved. Grid
Services was stable. The Service Challenges covered Tier-1 and Tier-2
testing, and there were sections on Security and Outlook.
TD noted that one part missing is the Tier-1 planning by JG. The
spreadsheet needs to be summarised in the Tier-1 Report - it is on the
Tier-1 Board page - and a note from JG is required on what Tier-1 will
provide in 2007 in terms of MoU. It was agreed that a section on this
would be added. It was important to convey the planning for 2007 and the
purchases for 2008. There was a discussion on whether to include this and
how it should be presented. No financial information required to be
included just the plan on how to meet the requirements, ie: the pledges in
JG's spreadsheet.
AS noted that BaBar planning should also be included.
TD also thought that highlighting occupancy for the year should be
emphasised. The Report numbering required to be reviewed and amended.
The words 'broadening of service challenge requirements' should be
re-phrased. Should Figure 4 be adusted?
h) Tier-2 Report
-----------------
SL presented his Report. Most of the information was in the Outturn
Report, but here he was looking at Tier-2 through experiment eyes. It was
noted that RJ still had to provide a paragraph on Tier-2 use from ATLAS.
DN still had to provide a paragraph on Tier-2 use from CMS.
TD noted that the following should be added in the Introduction: 'reflects
renewed effort to ensure the Tier-2s are seen as effective by each VO' -
how do they use Tier-2 and is it working? Performance Monitoring should
also be referred to in the Introduction. The Report needs to reflect how
are the Tier-2s working currently for each of the three large VOs? TD
noted that relating to ATLAS, the email summary would be good to use to
show a status summary. It was noted that the table requires description.
The deadline is Wednesday, the Executive Summary will be done by Thursday.
Other points were noted:
- in LHCb section, remove 'L' in second paragraph
- for Glasgow, change to 'this is now working'
- for London, remove 'when it is up and running'
- for SouthGrid it was felt better to mention all other SouthGrid sites
which are working before noting that there is a problem at Cambridge
- in the summary of sites and data, remove the figures after the decimal
points; two digits is enough
- an overall Summary should be added at the end stating that the situation
is not that great, but good enough, and that there is work onoing to
improve performance by measures discussed in the Performance Metrics
Report
- it was noted that a user perspective view should also be given - only
four sites are working properly, and that experiments are not taking
WLCG Tier-2s seriously - a measured statement is needed
- it was noted that inputs were required by Wednesday at the latest
i) Dissemination Report
------------------------
SP reported on events, the RTM, press releases and coverage, industry
projects, website, LHC@home and future work. TD noted that the words
'Since the last meeting' in the Introduction should be changed to read
'Since the last Dissemination Report' so that the report covered the year.
SP noted that the previous (intermediate) executive summary did have some
dissemination information. Any figures/information quoted should be for
the full year.
j) Documentation Report
------------------------
SB reported on the website, user area, and deployment document. One piece
of recent news was the release of the updated gLite User Guide:
https://edms.cern.ch/document/722398/1.1
SB further reported on the User Information Group, the external user
guide, and future work. TD noted that the links should be standardised
throughout. It was felt that a positive emphasis should be maintained by
a statement saying that the user guide is much improved, and any other
improvements over the year should also be drawn attention to. The OC will
note p2 at the end, therefore it was agreed to amend the wording to: 'will
ultimately depend on user feedback which has been limited so far'.
Under EGEE, the timescale should be checked with regard to the prototype
website, and say, rather, that: 'this is a recognised problem area'. PC
to tell NeSC that the timescales are slipping for the production website.
k) Performance Metrics Report
------------------------------
JC reported that the plots had been updated and he had added new plots
regarding CPU and disk. JC was asked to emphasise the 100% line in his
figures, and it was agreed that a bar would be added. For utilisation,
the variations are explained, however the metrics, what is available, plus
monitoring of efficiency numbers are still to be incorporated.
TD noted that there might be an overlap here with the Outturn Report and
SL's Tier-2 Report. This needs to be rationalised to ensure we aren't
contradicting ourselves. In Figure 11 it was noted that the scheduled
downtime is a combination of the colours. There was a discussion as to
whether the numbers could be presented differently and percentages given,
ie: a filtered statement rather than the graph. It was also noted that
the information here contradicted SL's plots. Was the hardware delivered
to GridPP the same? Was Figure 3 consistent with SL's ratios? It was
noted that the promised delivered ratio was not the same. Edinburgh was
at 152 in SL's table and 300+ in JC's table. It was reported that JC was
providing a figure for the final quarter whereas SL was providing for the
whole year. It was agreed that JC/SL would iterate and cross-check.
It was noted that the difference could be between promised/delivered for
CPU and storage. JC should use the same definitions as SL in the Outturn
Report, viz: promised, delivered, available, used.
l) 2006 Outturn Report
-----------------------
(note that disk outturn for 06Q4 is currently being reviewed)
SL reported on definitions, Table 1. The Tier-2s have met CPU and disk
about 50% - resource usage fluctuates as does use-to-ratio. TD noted that
the OC would interpret this as utilisation of Tier-2s being bad - but this
relates to the experiments. It was noted there was Tier-1 inconsistency.
SL reported that Table 3 was incorrect but this had now been rectified.
The Tier-2 report could cross-reference this table and put it in context.
It was noted that the disk table was incomplete and innacurate - the table
could be re-constructed to give a snapshot of now. This could include its
three Quarters historically with Quarter 4 being 'current status' - plus
some explanatory text. For table 1 there was a RAL PPD and Tier 1
question with reference to 50 TB of loaned disk. The User Board figures
from CMS will help. All available details should be emailed to SL.
2. Risk Register Review
========================
It was noted that the current Risk Register had been circulated by email.
a) LCG
-------
TC noted that R22 should be changed to 2 and 3 respectively. R25 should
be altered to 2 (and 2). R10 and R38 should remain as is.
b) MSN
-------
RM noted that regarding R5, he was unsure of making changes. For R16, the
impact 2 should remain unchanged. R40 should be altered to 3 (and 3).
In the light of GridPP3 confirmation, this may have to be altered again to
4 (and 3).
c) Applications
----------------
RJ noted that R40 should be raised to 4 (and 3). R15 should be changed to
3 and 3. R22 should drop to 2 (and 3).
d) Production-Grid
-------------------
JC noted that R22 should be reduced to 6 months instead of 9. R32 should
be reduced to 2 (and 3). Current risk register forward earlier.
3. Overlaps and Timescales for completion
==========================================
It was noted that DB had not yet produced a ProjectMap Report or a
Resource Report. DB needs inputs from various individuals regarding the
ProjectMaP. Time constraints were understood.
The meeting concluded that all next versions should be sent to PMB marked
"final" by Tuesday (if possible) and Wednesday, noon (if required) for
upload to the OC page.
Exceptions:
* Deployment Report (DK)
* MSN Report (RM)
* Applications Report (RJ)
* ProjectMap Report (DB)
* Resource Report (DB)
- drafts should be circulated as soon as practical (latest Wednesday
noon). The executive summary draft will be circulated on Wednesday
evening (based on all inputs received on Wednesday) for comments by noon
Thursday.
[Note: all documents were made available to the OC by Thursday's deadline]
As the meeting was timed-out, TD noted that weekly reports will be minuted
and actions also, but no discussion was possible this week.
4. AOCB
========
Other items should be reviewed next week if possible.
STANDING ITEMS
==============
SI-1 Dissemination Officer's Report
-----------------------------------
Two news items had been published this week, on the LHCb software course
and EUGridPMA meetings. Thanks to Karl Harrison and Dave Kelsey for their
work on these. The PMA article had been picked up by GridToday. A GridPP
article on the Glue schema is due to be published in iSGTW this week, and
thanks to Stephen Burke for his help with this. There will also be a press
release this week about the latest WISDOM data challenge. SP is expecting
information from AS on the Tier-1 upgrade, JC on the wLCG workshop and DN
on the CMS workshop.
SI-2 Tier-1 Manager's Report
----------------------------
Hardware
========
1) Supplier One delivery
Most of the disk servers were handed off to the CASTOR and dCache teams
over a week ago (no report last Monday). One (of two) is now available in
dcache for ATLAS, the second will be deployed shortly. The servers
schedule for CASTOR (14 initially) are being "CASTORised" and discussions
are also underway with the experiments as to how the space should be
configured in CASTOR.
2) Supplier Two Delivery (I)
Acceptance testing continues and is on track to complete on time. We have
no significant problems at this stage that will delay acceptance in mid
February.
3) Supplier Two Delivery (II)
This delivery arrived on schedule and 36 (of 39) servers completed
supplier load test and have been handed over to us. The remaining 3 needed
memory card replacements and have recommenced supplier tests. Our 28 day
load test will commence later this week and will then run for 28 days.
4) Tape Purchase
Approval was given at the Tier-1 board to purchase an additional 350TB of
tape (in addition to the previous 200TB). A tender closes this Wednesday
and an order will be placed shortly afterwards.
5) Tape drive purchase
Approval was given at the Tier-1 board to purchase 3 additional tape
drives (and associated servers). A tender for this has completed (was
already running when approval received) and we expect to order shortly.
Lead time to delivery will be 2-3 weeks from placing an order.
6) Tape drive servers
A purchase will be made for the matching 3 tape servers to go with the
newly ordered drives.
Service
=======
No major issues, but continuing problems with:
- BDII availability (we plan to add a second BDII this week)
- dCache reliability (we will commence work on a new release this week)
- RB availability - under investigation
Staff
=====
1) James Adams started at RAL today as hardware technician
2) A new member of staff is scheduled to start on 1st February
SI-3 Production Manager's Report
--------------------------------
JC reported that most of the deployment team members and UK sysadmins were
at CERN last week for the WLCG collaboration workshop therefore he had few
items to raise this week. He will prepare a summary of areas covered at
the workshop that impact the UK.
1) GridPP was very well represented at the workshop and after CERN was the
biggest contributor. Many of our group led sessions or presented giving
a good impression of GridPP.
2) With several sites increasing their resources and after resolving a few
issues (e.g. LHCb at Glasgow) we have been seeing utilisation
maintained at the 70-80% level for the last week.
JC noted that he had OS problems on his laptop today and may not be able
to join the meeting, he therefore provided updates on the actions against
him (these are noted below).
SI-4 LCG Management Board Report
--------------------------------
See https://twiki.cern.ch/twiki/bin/view/LCG/MbMeetingsMinutes
SI-5 Documentation Officer's Report
-----------------------------------
SB reported the release of the updated gLite User Guide:
https://edms.cern.ch/document/722398/1.1
REVIEW OF ACTIONS
=================
236.4 DN to summarise and circulate the CMS model as a basis for discussion.
236.6 GP to summarise and circulate the LHCb model as a basis for discussion.
240.3 JC to remind sites to build-up disk resources in order to cope with
spring/summer '07 requirements.
This was done at the UKI meeting on 17th and via direct reminders. Item closed.
241.5 JC to circulate an email to the PMB regarding users' CPU time
requirements: guidance on user time, job requirements, and memory was
required, also information on default behaviour - this information then to
be relayed to tutorials by way of Experiment Representatives (RJ, DN, GP).
JC has material from SB and David Colling and will circulate it
shortly. SB informed JC that the user guide has been updated. JC also
mentioned the issue during his WLCG talk for the UK region. Item closed.
242.1 RM to summarise the meeting notes of the UK & Ireland Federation F2F
Meeting and circulate to the PMB.
[done, as discussed at F2F] Item closed.
242.2 DK to send a draft news item on the PMA meeting to SP.
[done, final news appeared] Item closed.
242.3 JC to volunteer someone from the DTeam to write a draft news article
on the WLCG workshop, to go to SP. We will be discussing outcomes at the
DTEAM meeting tomorrow. Item closed.
242.4 SP to email Hannah Cumming at TOTAL regarding a news item when the
VO goes live. [done, SP awaiting a reply]. Item closed.
242.5 SP to forward information regarding the call for abstracts for the
EGEE User Forum in May. [done]. Item closed
242.6 JC to raise the manpower issue at the Tier-2 Board regarding 'Winter
Holidays' and staff not being available in January to bring sites back
online. Co-ordinators are to investigate this.
Awaiting confirmation of the next T2B date. Item closed.
242.7 SP to email UKHEPGRID regarding citing of eScience papers.
[done following the meeting] Item closed.
242.8 Following the CMS Grid workshop at Bristol, DN to write a news item
regarding CMS, LHCb, and ATLAS.
[LHCb news has appeared. CMS one to appear? Should probably refer to ATLAS
workshop this week for future news]
Item closed.
242.9 SS to update and upload the next listing of PMB documents for the OC
committee.
[done following the meeting] Item closed.
244.1 DB noted that a statement needs to be made regarding applications
being cut, however an approach needs to be made to PPARC regarding this.
[done following the meeting - incorporated into exec summary]
Item closed.
244.3 JC/SL to iterate and cross-check with reference to the tables in
their respective reports.
[done following the meeting] Item closed.
ACTIONS AS AT 29.01.07
======================
236.4 DN to summarise and circulate the CMS model as a basis for discussion.
236.6 GP to summarise and circulate the LHCb model as a basis for discussion.
Next meeting will be on Monday 5th February. The meeting closed at 3.30 pm.
|
