Dear All,
Please find attached the GridPP Project Management Board
Meeting minutes for the 509th to 511th meetings.
The latest minutes can be 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, Dave.
GridPP PMB Minutes 509 (04.11.2013)
===================================
Present: Dave Britton (Chair), Pete Gronbech, Andrew Sansum, Jeremy Coles, Steve Lloyd, Roger Jones, Dave Colling (Minutes - Suzanne Scott)
Apologies: Dave Kelsey, Pete Clarke, Claire Devereux, Tony Cass, Tony Doyle
1. Oversight Committee (OC) Meeting
====================================
DB had circulated a summary of this meeting, which had taken place on Friday 1st November. No issues had been raised by the Committee regarding GridPP4 - congratulations had been voiced on a successful execution of the Project. They were focused on the strategy suggested for going forward, including the spend on the EGI subscription and tape media for GridPP5, bearing in mind the underspend at RAL. We would need to check the profile to determine the actual amount of the underspend. Approval in principle had been given to spend this.
It was apparent that the outlook for GridPP5 could be problematic. The issue is that 'flat cash' is probably the most optimistic scenario, which would already require a 10% reduction in real terms, and significantly less optimistic scenarios would be very challenging.
Actual figures were not known at this point. DB would talk further with Tony Medland. The OC were being supportive in trying to reduce the PPRP process. GridPP could begin writing its inputs to the Strategic Review immediately. The outcome from the Programmatic Review (PR) was known internally at STFC.
GridPP must now write strategic documents and address some possible scenarios along with strategic consequences of any funding decision. DB noted we also must include some 'forward' issues like IPV6 and CASTOR, which need to be addressed in the longer term regarding the evolution of infrastructure. DB noted that various documents would be required, including:
1. Experiment requirements (including hardware);
2. Tier-1;
3. Tier-2;
4. Strategic case for deployment/ops/support.
A case also needed to be written for each post and how to deliver requirements in different scenarios, including management. These documents would form the basis of a proposal; this had to be completed by January 2014.
It was noted that the Oversight Committee had been very supportive. Mike Whalley was stepping down. JC asked about EGI and international issues? DB agreed yes, this was an important area but currently ill-defined. We did need to address it. DB advised that eventually, once the landscape is known, we had to go to the Collaboration Board (CB) and ask questions about direction and the choices which had to be made. We would take this process step-by-step. DB hoped to speak to TM this week.
2. Network Forward Look
========================
It was reported that PC had dinner with Jeremy Yates recently, and discussions were ongoing regarding joint GridPP/DIRAC work. The PMB endorsed this concept of collaborative working.
3. AOCB
========
- JeS forms & timing for some sites
PG reported that DC had spoken with UCL and DB with Durham. DC would contact the appropriate people again today. DB had had an exchange with PC, who was the current Chair of ScotGrid, and it was felt to be a good idea to delay the Durham grant until next year. Durham would be advised that they would not receive any Tier-2 hardware grant at present.
PG noted that he had only seen a few grants so far on JeS. PG/DB would iterate regarding the GridPP pledges for CPU and disk at each institute.
STANDING ITEMS
==============
SI-0 Development Group Report
------------------------------
DC reported that he hadn't held the last meeting; one was due. He would circulate an email update.
SI-1 Dissemination Report
--------------------------
The new Dissemination Officer, Tom Whyntie, reported as follows via SL:
CHEP2013 - 14th - 18th October 2013
-----------------------------------
Amsterdam may have been a bit gray and wet, but the Beurs van Berlage was a magnificent venue for the 20th International Conference on Computing in High Energy and Nuclear Physics, CHEP2013. GridPP was strongly represented with the following poster and presentation
contributions:
* VomsSnooper - a tool for managing VOMS records
* Simulation and analysis of the LUCID experiment in the Low Earth Orbit radiation environment
* Use of VMWare for providing cloud infrastructure for the Grid
* CernVM-FS - Beyond LHC Computing
* Running jobs in the Vacuum
* DIRAC Distributed Computing Services
* MICE Data Handling on the Grid
* ATLAS Cloud Computing R&D
* T2K-ND280 Computing Model
* A Voyage to Arcturus
ScotGrid's "Voyage to Arcturus" poster won the chance to give a "Lightning Talk" on the Friday morning; Mark Mitchell subsequently gave the talk on Glasgow's efforts to standardise the build model for a Tier-2 site.
Further information:
* [CHEP2013 homepage](http://www.chep2013.org/)
* [A Voyage to Arcturus - CHEP2013
poster/slides](https://indico.cern.ch/contributionDisplay.py?confId=214784&contribId=1)
New GridPP Dissemination/Impact Officer - October 2013
------------------------------------------------------
At the end of October 2013, Tom Whyntie started in his new role as Dissemination/Impact Officer for GridPP, replacing Neason O'Neill.
Formerly of the CMS experiment (through Imperial College London), Tom has spent the past year as Researcher in Residence for the CERN@school project. Funded by a Science and Society Large Award, CERN@school has recently expanded into a national educational and research programme that provides students and teachers with pixel detectors developed by the Medipix Collaboration. The Langton Ultimate Cosmic ray Intensity Detector experiment (LUCID) is due to launch aboard Surrey Satellite Technology Limited's (SSTL's) TechDemoSat-1 in 2014, which will also provide school students with the opportunity to analyse data from the Low Earth Orbit space radiation environment. By formally becoming part of the GridPP team, based at Queen Mary, University of London, Tom will be working to combine the two roles by developing projects such as LUCID that bring together schools, academic institutions and industry through the use of GridPP; see, for example, the talk on LUCID at CHEP2013 (link below).
Further information:
* [The CERN@school website](http://cernatschool.web.cern.ch/);
* [CHEP2013 conference presentation about LUCID simulations](http://indico.cern.ch/getFile.py/access?contribId=415&sessionId=3&resId=0&materialId=slides&confId=214784);
* [TechDemoSat-1 mission page](http://www.sstl.co.uk/Missions/TechDemoSat-1).
Dr Tom Whyntie
GridPP Dissemination Officer, Particle Physics Research Centre School of Physics and Astronomy Office 609A, G. O. Jones Building Queen Mary, University of London Mile End Road, London E1 4NS
e-Infrastructures for SMEs - 30th October 2013
----------------------------------------------
On Wednesday 30th October, Prof. David Britton and Tom Whyntie attended a meeting hosted by HPC Wales at the Institute of Civil Engineers to look at how the High Performance Computing centres and innovation Catapults can best engage with Small and Medium Enterprises (SMEs). After hearing status reports from the UK's HPC Centres, fruitful discussions were had about the successes and challenges of engaging with SMEs - and the overwhelming message was that people, training, and ease of use were key factors. There may well be support for these kinds of resources and activities from the E-infrastructure Leadership Council (ELC).
SI-2 ATLAS weekly review & plans
---------------------------------
RJ reported that RAL power was due to go off tomorrow. ATLAS were giving up on UCL as the site needed too much work for it to remain useful, and it was very problematic for the ADC so they were withdrawing. UCL simply wasn't delivering and there was too much effort being poured into it at present. DB considered it was a question of it being a Tier-2 or a Tier-3 - was UCL no use as a Tier-2? RJ confirmed this was correct - there was no worker node; they had not migrated to SL6; they were too far behind now and taking up too much of people's time during shifts. DB asked RJ to write to him formally regarding ATLAS no longer wishing to use UCL as a Tier-2 site.
ACTION
509.1 RJ to write to DB regarding ATLAS no longer wishing to use UCL as a Tier-2 site.
DB advised that the funding they receive is for supporting ATLAS, so it needed to be discussed whether or not the agreement was a lost cause at this point, and if so, what the financial implications were.
SI-3 CMS weekly review & plans
-------------------------------
DC advised there wasn't much to report - some cloud work was progressing; things were running smoothly.
SI-4 LHCb weekly review & plans
--------------------------------
PC was absent.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) The October WLCG T2 availability/reliability figures have been made available: http://sam-reports.web.cern.ch/sam-reports/2013/201310/wlcg/WLCG_Tier2_OPS_Oct2013.pdf. Six GridPP sites fall below the 90% target in one or more areas – several of these are due to SL6 downtimes. I will report in more detail at the next PMB meeting.
2) The majority of GridPP sites have migrated to SL6. We do however have two sites that still have (more than 30% CPU capacity on) SL5 WNs. One faces manpower issues and the other machine room maintenance. Our deployment focus is now shifting to glexec enablement.
3) At tomorrow’s ops meeting we will discuss the WLCG approach to middleware validation. I expect work of the WLCG operations coordination task force in this area to pick up in the coming weeks, so direct site admin input at this time may be very useful. At a minimum it will ensure concerns are aired and documented.
4) UK eScience CA infrastructure services are at risk today, tomorrow (4th and 5th November) and possibly on 6th November due to electrical work.
For information:
A) The agenda for the WLCG workshop (11th-12th November) next week is now complete: https://indico.cern.ch/conferenceOtherViews.py?view=standard&confId=251191
B) EGI-InSPIRE is planning a workshop on the topic “EGI towards Horizon 2020” (see go.egi.eu/h2020 for more information).
SI-6 Tier-1 Manager's Report
-----------------------------
AS reported as follows:
Fabric
--------
1) Currently in standstill for disk and CPU tenders
2) Two T10KD tape drives ordered - waiting for delivery.
3) Machine room UPS power down scheduled for tomorrow (Tuesday 5th and Wednesday 6th). Intervention to upgrade essential power board supply and carry out periodic safety checks. We expect only a short break in service for National and international services such as BDIIs. Expect CASTOR to be down for a long working day, but back up late afternoon/evening on the 5th. Batch farm may be resumed same day or may run on until the Wednesday. Entries are in GOCDB for exact times.
Service
---------
1) Reports covering last week available at:
http://www.gridpp.ac.uk/wiki/Tier1_Operations_Report_2013-10-30
2) CASTOR
3) We expect to move all our batch capacity to Condor after the end of the shutdown
4) AFS service was switched off on 1st November
5) CVMFS client version 2.1.15-1 has been rolled out to all worker nodes in the Condor farm
SI-7 LCG Management Board Report
---------------------------------
There had been no meeting.
REVIEW OF ACTIONS
=================
496.2 PC to update the network forward-look. Ongoing.
500.3 AS to send details to DB regarding the RAL Tier-1 kit available for retirement, to enable DB to write to institutes, following which we would decide how to proceed. Ongoing.
508.1 DB to write to Manchester and make the salient points regarding the necessity of early procurement. Done, item closed.
ACTIONS AS OF 04.11.13
======================
496.2 PC to update the network forward-look.
500.3 AS to send details to DB regarding the RAL Tier-1 kit available for retirement, to enable DB to write to institutes, following which we would decide how to proceed.
509.1 RJ to write to DB regarding ATLAS no longer wishing to use UCL as a Tier-2 site.
Next PMBs
=========
18th November
21st November (F2F @ Imperial)
GridPP PMB F2F Minutes 510 (21.11.2013)
=======================================
Held at Imperial College, London
Present: Dave Britton (Chair), Pete Gronbech, Andrew Sansum, Dave Colling, Steve Lloyd, Pete Clarke, Claire Devereux, Roger Jones, Dave Kelsey, Tony Doyle, Tony Cass
Apologies: Jeremy Coles
GridPP5
=======
This PMB F2F meeting was held in order to discuss the GridPP5 proposal.
Due to the uncertainty around the outcome of the Comprehensive Spending Review (CSR) and the Programmatic Review (PR), a wide range of financial scenarios were considered.
The purpose of the meeting was to outline a set of possible strategies that might have to be considered when the review outcomes were known. No decisions were made, although there was a consensus that:
(a) flat-cash scenarios, which are effectively a 10% reduction, could be managed with significant belt-tightening;
(b) there was not much space below a flat-cash settlement where GridPP could continue to do all the things that are currently done;
(c) beyond that, significant cuts to GridPP would result in a significant reduction to the scope of GridPP because it simply was not an option to reduce all-areas pro-rata and fail at everything.
There were no obvious, or acceptable, areas that could be lost. Nevertheless, once the outcomes of the reviews were known, GridPP would need to come to a consensus in order to provide input to a strategic review early next year.
GridPP PMB Minutes 511 (02.12.2013)
===================================
Present: Dave Britton (Chair), Pete Gronbech, Andrew Sansum, Jeremy Coles, Steve Lloyd, Tony Doyle (Minutes - Suzanne Scott)
Apologies: Roger Jones, Dave Kelsey, Tony Cass, Pete Clarke, Dave Colling, Claire Devereux
1. Feedback/Action from the Oversight Committee (OC)
=====================================================
A document from the OC had been circulated. There was minimal content - they thanked the PMB for the documents which had been provided; they acknowledged that GridPP had delivered the project well during GridPP4. Despite the support of the OC, for political reasons the RAL 'underspend' could not be spent on the EGI subscription but we could spend this on tape media meantime if possible. The OC had agreed that the GridPP5 process would be streamlined as much as possible, and a timetable provided.
It was agreed that AS/DK should do the outturn forecast, look at the possible spend on tape media and advise Tony Medland about the profile for next year. A realistic outturn forecast for travel was also required.
ACTION
511.1 AS/DK to do the outturn forecast, look at the possible spend on tape media and advise Tony Medland about the profile for next year. A realistic outturn forecast for travel was also required.
2. GridPP5
===========
It was reported that, if required, a Collaboration Board (CB) meeting had been scheduled for 17 December at 11.00 am. It was noted that the outcome of the Programmatic Review (PR) was still awaited. The Strategic Review (SR) document was required by mid-January 2014. It was agreed that if there was no funding announcement by STFC before the Christmas break then it would be very difficult to provide the CB with any useful information. In this case it would be preferable to delay the CB until more information was known.
In the meantime, there were some things we could do to progress the GridPP5 proposal:
1. AS should look at the possible scenarios and strategy re the Tier-1;
2. SL should look at the leverage in relation to the Tier-2, including power issues, and this information would be required for the SR. SL would remind Institutes that the information was urgently required.
3. DB would begin to write inputs for the SR, and an earlier meeting would be preferable, we should push for this if possible.
3. Distributed Competence Centre
=================================
An email had been circulated. DB considered that this was an NGI issue - there were virtual teams existing across the EGI user communities to help uptake of EGI services. We did some of this at institute level and it was also a gateway to Horizon2020 funding. The UK NGI should respond to this and say that as a community we are interested. GridPP could be involved in the DCC as it would be useful when we talked to SMEs in the UK.
ACTION
511.2 CD to discuss GridPP's input with the UK NGI concerning interest in the Distributed Competence Centre.
4. AOCB
========
PG had put forward a proposal for a HEPSYSMAN one-day meeting in January 2014. This would be of minimal cost financially. There had already been offers of talks and interest expressed in such a meeting. Birmingham had offered to host. This was agreed, PG should go ahead and organise it.
STANDING ITEMS
==============
SI-0 Monthly Development Group Report
--------------------------------------
There was no report.
SI-1 Dissemination Report
--------------------------
Tom Whyntie reported via SL as follows:
1. cernatschool.org VO
* Working with Chris Walker (QMUL) to test the backup VOMS servers for small VOs - will be required for SMEs joining through small VO mechanism. Updating and feeding back on the Wiki page for all GridPP.
* Catalin Condurache (RAL) has set up CVMFS access for cernatschool.org - the next step is to deploy and test CERN@school-specific software; feedback and support from Matthew Doidge and CW helping enormously.
Useful links:
[https://www.gridpp.ac.uk/wiki/Adoption_of_Backup_GridPP_Voms_Servers]
(Backup VOMS servers on the GridPP wiki);
[https://ggus.eu/ws/ticket_info.php?ticket=98122](GGUS ticket for cernatschool.org CVMFS roll-out).
2. TSB competitions
* Alex Efimov is preparing a webinar with the Technology Strategy Board to advertise collaboration opportunities for GridPP and SMEs for two upcoming competitions:
** "Data Exploration" - £100k-£500k, 6-24 months (see link below);
** "Technology Inspired Innovation" - £33k for short-term projects with SMEs/micros;
Useful links:
[https://www.innovateuk.org/competition-display-page/-/asset_publisher
/RqEt2AKmEBhi/content/data-exploration-creating-new-insight-and-value?
p_p_auth=cRQMvjQ7](The Data Exploration TSB competition).
3. SME approaches:
TW has made approaches to the following SMEs:
* SSBV Aerospace & Technology: through a Uni. of Southampton SEPnet intern who presented his work on satellite payload testing at the Royal Institution SEPnet student expo on 20/11/2013;
* Palantir Technologies (UK office): through Dr Alex Sparrow (former Imperial/CMS physicist, now at Palantir);
* PythonAnywhere: cloud-powered, browser-based Python environments - for educational purposes (c.f. CERN@school coding/data analysis). Lunchtime meeting arranged 3/12/2013 with the dev team at their Clerkenwell Road offices.
SI-2 ATLAS weekly review & plans
---------------------------------
RJ was absent.
SI-3 CMS weekly review & plans
-------------------------------
DC was absent.
SI-4 LHCb weekly review & plans
--------------------------------
PC was absent.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) The UK CA was ready to move to SHA-2 certificates by default on 1st December, but it was decided to postpone until the start of January given that WLCG was largely but not completely ready for the switch over. (The GridPP website is not quite ready for SHA-2 as the general server migration/upgrade is not yet completed – due to wiki/database issues).
2) EGI has indicated a deadline of the end of December for NGIs to have a national ARGUS server – we are seeking clarification on what this means in practice. For the UK the central ARGUS instance will be hosted at RAL.
3) Glexec has become an ops critical test for those sites that indicate enablement/support. This has led to problems at a couple of sites but is now in hand.
4) We have made progress in several areas in support of our smaller VOs.
Firstly the regional Nagios has been configured to submit jobs for several of these VOs to enable dashboard monitoring of them.
Secondly DIRAC has been setup at Imperial to test work for T2K, landslides and NA62. More experience needs to be gained by our sysadmins (so the GridPP VO will be enabled next) and in using more complicated end-to-end needs of SNO+.
Thirdly sites are ready for use of the backup VOMS infrastructure – though ideally the smaller VOs will test it before we put it fully into production.
Fourthly, a tool to help with the deployment of new UIs has been made available to sysadmins/users: https://www.gridpp.ac.uk/wiki/User_Interface_%28UI%29_to_support_approved_VOs.
And finally, LFC webdav now appears to be working and supported at RAL.
5) We are currently reviewing GOCDB NGI associations and permissions.
6) A new WLCG core operations task force has been setup to investigate and coordinate multi-core work. It was requested ATLAS and CMS with LHCb also expressing interest. Alessandra will chair the group which has the mandate defined at: https://twiki.cern.ch/twiki/bin/view/LCG/DeployMultiCore.
For information:
A) There is a pre-GDB on 10th December on Identity federation in WLCG: https://indico.cern.ch/conferenceDisplay.py?confId=272770. The agenda for the GDB on 11th December is at https://indico.cern.ch/conferenceDisplay.py?confId=251192.
SI-6 Tier-1 Manager's Report
-----------------------------
AS reported that:
- Disk and CPU procurements were running on track and should deliver early in the new year.
- T10KD tape drives were delivered.
- Smaller purchases were now commencing.
- The new Condor batch farm was working very well - a huge improvement over torque/maui.
SI-7 LCG Management Board Report
---------------------------------
DB reported that a meeting had taken place on 19th November, but there had been no UK implications in any of the reports.
REVIEW OF ACTIONS
=================
496.2 PC to update the network forward-look. Ongoing.
500.3 AS to send details to DB regarding the RAL Tier-1 kit available for retirement, to enable DB to write to institutes, following which we would decide how to proceed. Done, item closed.
509.1 RJ to write to DB regarding ATLAS no longer wishing to use UCL as a Tier-2 site. Done, item closed.
ACTIONS AS OF 02.12.13
======================
496.2 PC to update the network forward-look.
511.1 AS/DK to do the outturn forecast, look at the possible spend on tape media and advise Tony Medland about the profile for next year. A realistic outturn forecast for travel was also required.
511.2 CD to discuss GridPP's input with the UK NGI concerning interest in the Distributed Competence Centre.
AOCB
====
PG reminded everyone about the Quarterly Reports for Q3 - these were now urgent. Could those who have not yet responded, do so immediately.
Next meetings:
--------------
- Monday 16th December @ 12 noon
- Monday 6th January 2014 @ 12.55 pm
|
