Dear All,
Please find attached the GridPP Project Management Board
Meeting minutes for the 557th to 562nd meeting.
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 Meeting 562 (20.04.15)
=================================
Present: Pete Gronbech (Chair), Andrew Sansum, Dave Britton, Dave Colling, Tony Doyle, Roger Jones, Tony Cass (Minutes Ð Suzanne Scott)
Apologies: Dave Kelsey, Steve Lloyd, Jeremy Coles, Pete Clarke, Claire Devereux
1. GridPP5 Feedback to Applicants
==================================
DB had received feedback from STFC regarding the GridPP5 Proposal, and had also received two Referee Reports. Specific questions had been asked to which the PMB were required to respond by Monday 11 May. The PMB could also respond to the Referee Reports, and it was noted that there were many things therein that we agreed with.
DB suggested that the PMB meeting go through the feedback and assign actions for draft responses. The discussion should be constrained to 1 hour.
The Referee feedback seemed reasonable and knowledgeable Ð the PMB did not know who the Referees were. The Committee Questions needed to be answered first:
Q1 Ð this was on the Tier-1 and Tier-2 threshold.
------------------------------------------------
If we argued based on strategy then we could justify this. All we could do between the 90% and 70% scenarios would be to cut the support for 'other' experiments. This would mean 10% less hardware funding and the loss of up to one FTE. It was agreed to explore this. WLCG activity had to be kept robust as a priority.
ACTION
562.1 DB to take ownership of Committee Q1 and prepare a preliminary draft response.
Q2 Ð this related to additional hardware greater than pledge, to ATLAS and CMS.
------------------------------------------------------------------------------
DB noted this was a Tier-2 hardware issue and we needed to explain the benefit of this and also what the costing is.
ACTION
562.2 RJ/DC to prepare, independently, a draft reply to Committee Q2, from their experiment's point of view, explaining what this 'additional' pledge was for and what the loss would be if this was not delivered.
ACTION
562.3 PG to work out the figure for the Tier-2s for CMS and ATLAS, with respect to Committee Q2.
Q3 Ð this was about Tier-2 facilities
-------------------------------------
DB considered that it was unreasonable to include leveraged resources in the pledge Ð no site could do this.
ACTION
562.4 SL/DB to draft a response to Committee Q3.
Q4 Ð this was about capping the UK contribution to WLCG
-------------------------------------------------------
DB considered that this issue was not something that GridPP could decide. The UK could not decide this unilaterally Ð it would have to be discussed and decided at higher level at the Resource Review Board (RRB). The consequence would be the same as reducing the trigger rate. RJ considered that the UK was an efficient and lean operation, we could devolve work from the Tier-1 to the Tier-2, GridPP was strong enough to be able to do this, however others were not.
ACTION
562.5 RJ to draft a response to Committee Q4.
Q5 Ð this was about the experiment support posts
------------------------------------------------
DB thought that we needed to give a robust reply to this question. The current structure is unique to the UK however there was a need in the UK for these posts.
We need to explain what these posts actually do, and this is not something that someone at CERN would do. What the posts do is also country-specific, the tasks performed are not done uniformly, internationally.
DB noted we should:
- mention the CMS people
- CMS considers that RAL is correctly aligned with experiment requirements
- GridPP2 failed to do this, and we consistently did not meet experiment expectations
AS noted that the core tasks could not be got rid of Ð they have to be done and it is better if they are done by someone embedded within the experiments.
ACTION
562.6 AS to prepare initial draft on Committee Q5, to explain the UK experiment support roles and why they are unique to the UK.
562.7 RJ/DC to add-in arguments to draft of Committee Q5, following AS, about the operational efficiency of the experiment support roles to the experiments.
Q6 Ð this focussed on use of new technologies
---------------------------------------------
6.1
We need to expand on this, as there was not enough space in the original Proposal to do so. Until, eg SKA, know what they want to do, it is very hard for us to be specific about interaction. AS noted there was also an STFC computing strategy but there was no information available on that yet Ð cost-sharing might come out of that. DC had a copy of the document which he would forward-on.
ACTION
562.8 AS/PC to draft an initial response to Committee Q6.1.
6.2
This had to be discussed with DC. We were trying to lead and push development in order to drive the direction of the experiments, to enable more effective and efficient support by GridPP, especially regarding testing and deployment, eg: multicore jobs, and the partnerships this involves. It was up to the sites to schedule work based on the resources available, eg: CVFMS.
ACTION
562.9 AS to draft an initial response to Committee Q6.2, and iterate with DC in order to complete this.
Q7 Ð this was on UKT0
---------------------
DB considered that this was an opportunity to expand the information presented in the Proposal, which had been curtailed due to a document size limit. We need to justify in more detail the opportunities of other funding etc.
DB noted that a 'shopping list' was required, a tiered list. It mirrors EUT0. There were opportunities here to express vision, and the focus should be on economies of scale.
ACTION
562.10 PC to draft an initial response to Committee Q7, getting inputs from AS as required.
Q8 Ð other sources of funding?
-----------------------------
RJ noted that we are a collaboration and in partnership with the experiments. We do software development only to the limited extent that this is funded by experiments.
AS noted that interaction with industry can be expressed easily here. RJ agreed, we liaised with vendors, do trials and tests at sites etc for vendors, therefore have direct contact with industry.
PG agreed, noting that vendors do take us seriously and fund our meetings; we also get preferential rates and discounts on kit because they benefit from our expertise.
DB noted that we could add cloud procurement using pre-commercial hardware.
ACTION
562.11 SL/DB/AS to draft a response to Committee Q8.
Q9 Ð Pathways to Impact
-----------------------
DB reiterated that we had to cut-out a large document relating to this from the original proposal due to document size limit. We would use what we had already available in the original document to answer this question fully.
ACTION
562.12 SL to draft a response to Committee Q9, utilising the original document we had available for the Proposal.
DB now turned to the Referee Reports which he had highlighted in green (where we agreed with the comments expressed) and red (where we had to respond to comments made).
Referee Report No 1
-------------------
page 2
Ð we agreed that Figure-1 was over-simplified, however this was deliberate as we were trying to express a very complex thing in as clear terms as possible. TD noted that we should use the comments made by Referees in order to help our response. DB thought we agreed with the viewpoint expressed on this page.
page 3
Ð we are linked to WLCG; we have quarterly meetings where the experiments are represented; there are weekly meetings as well to which they are invited; we allocate according to need.
Re the comment on 'isolated silos' Ð this assertion is not true. We treat the Tier-2s as an holistic system but we cannot fund fractions that are less than 0.5FTE. Sites support other sites: they comprise a service matrix that interlinks. AS noted that we could explain again about the challenges to get to 0.5FTE and why we are doing this. We have started the process already with the planned outlines and discussions at GridPP34. DB added that the scenarios ramp-down over a four-year period to get to 0.5FTE Ð it is therefore about system evolution. Small sites are supported by larger ones Ð they interlink.
DB noted that this entire large block of comment on page 3 required a response. TD advised that we minimise any new text and consistently refer back, where possible, to the arguments expressed within the Proposal.
ACTION
562.13 DB to draft a large paragraph/possibly page response to the block of comments from Refereee 1 on page 3, explaining strategy and relating this to the individual points made Ð referring back to the Proposal where possible.
page 4
- leverage: leveraged resources are in addition to the pledge; the risk is about exposure to electricity costs etc. DB noted we could add this to the Risk Register.
- network: the funding model needs to be explained Ð the network is 'top-sliced'
- we agree with some of the green-highlighted sentences here Ð we can quote these back in answers to questions.
- re the 'drastically different proposal' Ð we agree that in no circumstances should this be required. We don't agree that the direction as outlined is the correct direction to follow. It is unlikely that STFC would wish to do this. DB noted we should be prepared to answer this point of view, but it would not be helpful to include this in our response.
TD pointed out that this was an alternative proposal to a different call. It was not consistent with the brief requested.
- re the 'savings' generated by suppressing innovation, we agreed with the comment and agreed to echo this back in response to the question on Dirac etc. AS should include this.
- we don't agree that the funding scenarios for hardware 'could have been better'. We have documentation to back-up our strategy here.
ACTION
562.14 DB/AS to discuss and draft a response to the Referee 1 comment at the foot of page 4 on the funding scenarios relating to hardware.
page 5
- we agree with the comment that the Flat Cash scenario is the 'minimal needed' to achieve the project. We can re-use this comment.
- regarding reducing the pledge to WLCG, we can illustrate that this does not work. We support the intention of the Reviewer, but in practice are already at the threshold of being forced to reduce staffing levels.
- regarding the formatting of the document, if a word-count maximum had been applied by STFC (rather than a page count maximum) then we would have been happy to use bullet points, however we had to excise our bulleted lists due to the document being page-limited.
ACTION
562.15 DB to respond regarding the comment of Reviewer 1 on page 5 about formatting Ð STFC would have to change their requirements to a word-count maximum rather than a page-limited count, in order for us to be able to utilise different formatting.
Referee Report No 2
-------------------
DB noted that we agreed with many of the Referee statements, which DB highlighted in green, on pages 1, 2 and 3, and at the top of page 4.
Regarding the Referee comment on mitigation of risk in relation to funding scenarios, DB noted that funding scenarios were high risk, therefore there was no way of mitigating these. TD noted again that it would be useful to echo this back where possible.
Regarding the last paragraph of 'Additional Comments', it would be difficult to reply regarding CEPH, PRACE and EUDAT Ð our relationship with such things was via EUT0 and EGI. AS noted that the way in which these join together is complex and the Referee's misunderstanding of how it all works should be addressed.
TD considered we should also provide an update and take the opportunity to describe the evolving situation.
ACTION
562.16 AS to draft a response addressing Referee 2's misunderstanding (CEPH, PRACE, EUDAT) as expressed in the 'Additional Comments' on page 5.
Regarding ALICE, it should be noted that ALICE is dealt with by NPGP.
DB noted that initial drafts should be available by the end of the week, in preparation for being discussed at the PMB F2F at GridPP34 at QMUL.
2. GridPP34 Agenda
===================
PG advised that he had updated the Agenda based on last week's discussion and agreements. DB noted that Chairs needed to be assigned to sessions.
3. AOB
=======
- HEPSYSMAN: PG had circulated a request for funding of a 2-day meeting at RAL. Ian Collier would run the Security Workshop. The PMB agreed the request.
- OpenStack: Jeremy Yates had been in touch regarding an interest group on this. Would DC respond to the doodle poll? Was he attending? Yes to both.
DB noted that we might want to discuss OpenStack at GridPP34. Could this be raised in the discussion session? DC confirmed yes, there were additional discussion points he could include regarding clouds and how we progress these.
- UCL: JC raised the issue of UCL performance and the possibility of taking UCL offline and out of monitoring. DC reminded that UCL had sent a letter to us a year ago, advising that they wished to withdraw. DB noted that ATLAS had to tell us what they wanted to do. RJ reminded that ATLAS had already done so. DC proposed that UCL need to downgrade themselves from a Tier-2 to a Tier-3 and we could assist with this. This was agreed. PG would pass this onto the Ops Meeting and take technical steps to get UCL offline for the moment.
ACTION
562.17 PG to pass the PMB agreement to 'downgrade' UCL onto the Ops Meeting and take technical steps to get UCL offline for the moment.
- RAL 'Away Day': AS reported that SCD were having an 'away day' on Wednesday and Thursday of this week, at Telford. This will affect their responsiveness risk. AS advised that they would carry on with out-of-hours monitoring however with the issue of a potential hardware fault, they were 4 hours away. There was a low risk associated with this. They could not travel overnight therefore there was an increased responsiveness risk on Wednesday & Thursday.
DB re-iterated that the timetable for draft responses to the GridPP5 Questions and Reviews was by next week's F2F at GridPP34.
STANDING ITEMS
==============
Standing Items were not dealt with this time round due to time constraints.
REVIEW OF ACTIONS
=================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
561.1 PG to insert into the GridPP34 Agenda: statements from the experiments about the experiment vision through GridPP5 in the morning session.
561.2 JC to provide a bulleted list of items to be addressed within the structured Tier-2 discussion session at 3.30 pm.
561.3 PG to ask 6 people who went to CHEP to each give a 5-minute report in the 'Directions' session.
561.4 PG to ask Alessandra Forti to provide a short WLCG summary report.
ACTIONS AS OF 20.04.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
561.1 PG to insert into the GridPP34 Agenda: statements from the experiments about the experiment vision through GridPP5 in the morning session.
561.2 JC to provide a bulleted list of items to be addressed within the structured Tier-2 discussion session at 3.30 pm.
561.3 PG to ask 6 people who went to CHEP to each give a 5-minute report in the 'Directions' session.
561.4 PG to ask Alessandra Forti to provide a short WLCG summary report.
562.1 DB to take ownership of Committee Q1 and prepare a preliminary draft response.
562.2 RJ/DC to prepare, independently, a draft reply to Committee Q2, from their experiment's point of view, explaining what this 'additional' pledge was for and what the loss would be if this was not delivered.
562.3 PG to work out the figure for the Tier-2s for CMS and ATLAS, with respect to Committee Q2.
562.4 SL/DB to draft a response to Committee Q3.
562.5 RJ to draft a response to Committee Q4.
562.6 AS to prepare initial draft on Committee Q5, to explain the UK experiment support roles and why they are unique to the UK.
562.7 RJ/DC to add-in arguments to draft of Committee Q5, following AS, about the operational efficiency of the experiment support roles to the experiments.
562.8 AS/PC to draft an initial response to Committee Q6.1.
562.9 AS to draft an initial response to Committee Q6.2, and iterate with DC in order to complete this.
562.10 PC to draft an initial response to Committee Q7, getting inputs from AS as required.
562.11 SL/DB/AS to draft a response to Committee Q8.
562.12 SL to draft a response to Committee Q9, utilising the original document we had available for the Proposal.
562.13 DB to draft a large paragraph/possibly page response to the block of comments from Refereee 1 on page 3, explaining strategy and relating this to the individual points made Ð referring back to the Proposal where possible.
562.14 DB/AS to discuss and draft a response to the Referee 1 comment at the foot of page 4 on the funding scenarios relating to hardware.
562.15 DB to respond regarding the comment of Reviewer 1 on page 5 about formatting Ð STFC would have to change their requirements to a word-count maximum rather than a page-limited count, in order for us to be able to utilise different formatting.
562.16 AS to draft a response addressing Referee 2's misunderstanding (CEPH, PRACE, EUDAT) as expressed in the 'Additional Comments' on page 5.
562.17 PG to pass the PMB agreement to 'downgrade' UCL onto the Ops Meeting and take technical steps to get UCL offline for the moment.
The next PMB would be an F2F meeting and would take place on Tuesday 28 April at QMUL.
GridPP PMB Meeting 561 (13.04.15)
=================================
Present: Pete Gronbech (Chair), Steve Lloyd, Jeremy Coles, Pete Clarke, Andrew Sansum (Minutes Ð Suzanne Scott)
Apologies: Dave Britton, Dave Colling, Tony Doyle, Roger Jones, Dave Kelsey, Claire Devereux, Tony Cass
1. GridPP34 Agenda
===================
PG noted that he was awaiting inputs from DC and Andrew McNab on the 'evolution' session and also the structured discussion. PG asked if the refreshments and lunch timings were correct? SL did not know as Dan Traynor had booked these. PG should contact Dan direct to change the bookings if required.
There had been a talk suggestion from Alistair Dewhurst on 'monitoring' and possibly a talk from Glasgow on Graphite?
Was there a speaker for the opening session? The opening session was unfilled at present. SL suggested that Tom Whyntie could talk about Cern@School, the science aspects. AS asked for clarification about the main thrust of the Wednesday sessions? If these related to the evolution of GridPP5 then how would this vision be expressed?
SL considered that the experiments needed to tell us how they wanted to evolve. What had they done during the shutdown, and what was the forward look to GridPP5 and beyond? AS thought that if the evolution through GridPP5 was the focus then we definitely needed the experiment vision. The Tier-2 vision should be expressed as well. A clear vision for GridPP5 overall was required, as things have moved on from the status and situation of GridPP3 and GridPP4. SL added that there were large and small site agendas. JC noted that VAC & Vcycle were important for the smaller sites.
It was agreed that PG insert statements from the experiments about the experiment vision through GridPP5.
ACTION
561.1 PG to insert into the GridPP34 Agenda: statements from the experiments about the experiment vision through GridPP5 in the morning session.
AS considered that we needed a number of short talks to prompt discussion Ð VAC/Vcycle were good examples. SL suggested xrootd? Also a 'toolkit' talk from Tom Whyntie?
JC suggested the Tier-2 issue of service data centres Ð there was a lot happening in Institutes at present regarding central services and how resources are shared with the centre, including the issues involved.
Further discussion ensued. It was agreed to add-in the issue of Tier-2 evolution tools in an afternoon session; leave-out the 'guest' talk; and start the experiments' vision talks in the morning session.
The morning session would therefore commence with a welcome, followed by 20-minute experiment 'forward looks' which should set-the-scene for GridPP5. Talks from the 3 major experiments would be required, to include what they accomplished over Run-2 and/or a 5-year plan that would include GridPP5 Ð this would constitute 'experiment evolution in GridPP5.
AS considered that things would not be the same in 2019 and the experiments should focus on what was it that we were trying to build and achieve? SL noted we must give them a 'brief' so that they knew the focus Ð RJ, DC and Andrew McNab should speak.
Lunch would follow. Then the slot should focus on the Tier-2s. Should there be a talk on Ganga? Federated access? AS noted that the security model will change over the next 5 years: perhaps Jens could talk about authentication/authorisation at the end of the decade? including federated identity?
For the Tier-1 session, AS would talk about the changes in GridPP5, cost sharing etc, and he could cover EUT0/UKT0 as well. Would there be an LSST talk?
Re the discussion session at 3.30 Ð would this work? JC considered there was quite a lot to talk about. Could JC provide a bulleted list of items to address? Real structure was required for this session to ensure it was chaired effectively.
ACTION
561.2 JC to provide a bulleted list of items to be addressed within the structured Tier-2 discussion session at 3.30 pm.
AS would give a 10-minute talk plus 5-minutes for questions on the Tier-2 before the presentation on the HNSciCloud:H2020 proposal.
On Thursday there was a random 'directions' slot Ð could reports go here? We could ask the delegates who attended CHEP to give a 5-minute brief each, to express the different interests, to ensure the focus was not too narrow. A highlights talk. This would work for the WLCG workshop as well Ð we could ask Alessandra Forti to do a summary talk.
ACTION
561.3 PG to ask 6 people who went to CHEP to each give a 5-minute report in the 'Directions' session.
561.4 PG to ask Alessandra Forti to provide a short WLCG summary report.
It was agreed to move the CHEP/WLCG sessions to after lunch Ð 6 people should give 5-minute talks each.
PG would re-do the Agenda and upload the new version for further iteration.
2. AOB
=======
- AS had received an email from Jeremy Yates saying that STFC had funded his bid for 'disaster recovery' project for Dirac. It may be double-counted with respect to our bid? He would forward more information once it was available.
- An email had been received advising the PMB that the next GridPP Oversight Committee meeting would be on 9 November. No information or feedback about the GridPP5 proposal had yet been received from STFC.
STANDING ITEMS
==============
SI-0 Monthly Report from Development (Cloud) Group
---------------------------------------------------
DC was absent.
SI-1 Dissemination Report
--------------------------
There was no report.
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 noted nothing major to report.
SI-5 Production Manager's Report
---------------------------------
JC reported that ATLAS had been running a large number of simultaneous jobs. WLCG had issued the March reports for the Tier-2s. An outage had affected some services.
SI-6 Tier-1 Manager's Report
-----------------------------
AS had nothing major to report. The hardware arrived at the end of the FY as expected Ð he still had the expenditure outturn report to prepare. An offer had been made to a candidate for the hardware post. The sysadmin post had new shortlisted candidates. There had been a network outage and a Post Mortem would be prepared. The outage lasted over 12 hours. It related to a CASTOR upgrade which had been scheduled, prior to which an intervention had been made to remove a network switch that was faulty, which in turn caused a network loop that fouled-up the CASTOR upgrade. This took time to solve and was a major outage.
AS reported that the end of the Cream CE Service at RAL would happen soon Ð this was due to be retired in May; AS would circulate a summary.
SI-7 LCG Management Board Report of Issues
-------------------------------------------
There was no report.
REVIEW OF ACTIONS
=================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
ACTIONS AS OF 13.04.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
561.1 PG to insert into the GridPP34 Agenda: statements from the experiments about the experiment vision through GridPP5 in the morning session.
561.2 JC to provide a bulleted list of items to be addressed within the structured Tier-2 discussion session at 3.30 pm.
561.3 PG to ask 6 people who went to CHEP to each give a 5-minute report in the 'Directions' session.
561.4 PG to ask Alessandra Forti to provide a short WLCG summary report.
The next PMB would take place on Monday 20 April at 12:55pm.
GridPP PMB Meeting 560 (30.03.15)
=================================
Present: Dave Britton (Chair), Andrew Sansum, Dave Colling, Jeremy Coles, Tony Doyle, Pete Clarke, Pete Gronbech, Roger Jones, Dave Kelsey, Claire Devereux (Minutes Ð Suzanne Scott)
Apologies: Tony Cass, Steve Lloyd
1. ATLAS Tape Request
======================
AS had circulated a resource request from ATLAS, asking for more tape resources. GridPP has allocated tape resource in FY15 to Dirac, and then there will not be any extra available at all in FY16. The Tier-1 cannot therefore meet the requirements as requested, as they are already fully allocated, however AS confirmed they would keep tracking the situation and other opportunities may arise. DB considered that we could not guarantee anything at present. If there were to be a crisis, then we could help, but we cannot increase our planned commitment at this point. AS would respond to ATLAS accordingly.
2. GridPP34 Agenda
===================
DB had circulated an email with a suggested Agenda outline. It was thought best to be as open as possible with the community regarding the GridPP strategic evolution proposed. DB would email the CB with an update on the recent Oversight and Strategic Review meetings. The CB should discuss the GridPP5 proposals with relevant parties. The PMB agreed to discuss GridPP5 at the forthcoming Collaboration Meeting. PC thought that any tables in the proposal should be considered as being in the public domain. DB will have details available in order to respond to questions on the day.
It was agreed that the meeting themes could therefore revolve around the evolution of the Tier-1 and Tier-2s in GridPP5. CD thought we could present an outline in advance to the meeting, in order to invite questions and discussion. DB would give a talk in Session 1 about Project Status and GridPP5. Was another talk scheduled here also?
Session 2 was about how the Tier-2s would evolve Ð followed by a discussion session on the Tier-2s. DC would ask Andrew McNab about possible talks.
For Day 2 Ð it was agreed to look at the evolution of the Tier-1. AS could do this and suggest component talks, to include CEPH and Cloud etc, as well as other groups.
DB had to leave the meeting at this point, PG then chaired the meeting.
PG noted that the suggestion was that Session 5 could look at 'Directions' of other things like CHEP, WLCG and EGI. A CHEP report-back would be very useful. JC had a list of people who had intended to go. We should ask if someone wished to give a report-back.
From the EGI side, CD was not sure this would be of interest to the GridPP audience? DK considered there was no harm in informing people of what was going on. AS suggested we include SKA?
For Session 6 Ð could this be on security or networking? We could look at the status of IPV6? DK thought a networking topic would be good. DK could give a Hepix overview? It was a bit early for the new Security Officer to speak? The new Security Officer could certainly give an overview of his fresh perspective coming into the community, and perhaps raise issues the he feels need attention? This was agreed, but a short talk only.
Were there any other issues to add? AS considered that it would be good to finish on a 'feelgood' talk. JC suggested experiment slots for short updates from them, which could give a forward look at the possible physics in Run-2? PG suggested an outside speaker from LSST? AS thought it wasn't a good slot for an invited speaker.
3. AOB
=======
HNSciCloud
----------
AS had circulated an email. Helix Nebula had been pushing to get sites on board to go part-funded with commercial cloud services. At short notice RAL had been asked to get involved. There had been no time to do calculations and the detail did not seem attractive as a proposition in a number of areas, so they had declined. But now a representative from UK Government G-cloud was involved, so we could participate at the 75k Euro level of contribution. The resource may prove to be useful but it did not intrinsically benefit GridPP.
Tier-1 outage
-------------
There had been a Tier-1 outage on 28 March at around 9.45 pm on the Friday. The machine in the database rack had blown a fuse, which in turn blew a fuse on the APC strip which was supplying power to the network switch, which was providing networking to the core system, which subsequently went down. They need to do some work to make that bit of the network resilient.
Jisc Assent(Moonshot)
---------------------
Jisc Assent (formerly Moonshot) was being released in pre-production, running on various platforms. There were some early adopters. It is an authentication-only mechanism. A Pilot Study was planned re security etc. The international scaleability was limited. It builds on Eduroam technology.
STANDING ITEMS
==============
SI-0 Monthly Report from Development (Cloud) Group
---------------------------------------------------
There was no report.
SI-1 Dissemination Report
--------------------------
SL was absent.
SI-2 ATLAS weekly review & plans
---------------------------------
RJ had nothing major to report.
SI-3 CMS weekly review & plans
-------------------------------
DC noted nothing major to report.
SI-4 LHCb weekly review & plans
--------------------------------
PC noted nothing major to report.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) On Friday 27th, RAL lost access to a number of database systems located within the RAL Tier1. All Tier1 Castor was stopped. Atlas Frontier and GOCDB were also affected. Services were restored by 00:30 on 28th March.
2) The EGI operations portal was not working for parts of last week and the previous one. This had knock on impacts for ROD and security monitoring Ð for example (low availability) alarms were not being generated and followed up.
3) We have a DPM collaboration meeting this Tuesday.
4) HEPiX in Oxford last week went very well. Pete and his team did a great job. GridPP was very well represented in the full agenda: https://indico.cern.ch/event/346931/timetable/#all.detailed.
For information:
A) There was a WLCG OMB last week: https://indico.egi.eu/indico/conferenceDisplay.py?confId=2374. Nothing significant to raise to the PMB, but the EGI-Engage talk may be of interest.
SI-6 Tier-1 Manager's Report
-----------------------------
AS reported as follows:
Fabric:
1) CPU procurements Ð delivered.
2) Disk procurement Ð delivered.
3) Router issue continues - escalated to Level 1 support who will attend on site after Easter. Refactoring the Tier-1 network to allow more frequent iterations and testing.
4) Failure on the Tier-1 network at 21:43 on Friday 27th March caused loss of connectivity to core databases. Cause was a failed power supply on a server which in turn tripped a supply fuse on a APC unit causing loss of power in turn to a critical network switch. We are reviewing resilience of some parts of the core network.
SI-7 LCG Management Board Report of Issues
-------------------------------------------
There was no meeting.
REVIEW OF ACTIONS
=================
All actions were ongoing.
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
ACTIONS AS OF 30.03.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
The next PMB would not take place on Monday 6th April as this was Easter Monday. The next PMB date would be advised as soon as possible, and was likely to be dictated by the response of the Strategic Review Committee.
GridPP PMB Meeting 559 (23.03.15)
=================================
Present: Dave Britton (Chair), Andrew Sansum, Dave Colling, Jeremy Coles, Tony Doyle, Pete Clarke, Steve Lloyd (Minutes Ð Suzanne Scott)
Apologies: Pete Gronbech, Roger Jones, Dave Kelsey, Tony Cass, Claire Devereux
1. Report from Oversight Committee (OC) Meeting
================================================
DB had circulated a brief report. The meeting had gone reasonably well Ð the Committee were happy with progress. Questions were raised about GridPP5, some of these were helpful in the context of the forthcoming GridPP Strategic Review. The Committee membership for the Strategic Review was now known. The OC were supportive of the 'flat cash' and '90%' plans presented by PMB members, and considered that the '70%' scenario would cause chaos.
2. GridPP Strategic Review
===========================
This Review meeting would take place on 24th March. DB was working on a draft presentation that incorporated information presented to the Collaboration Board and to the Oversight Committee. Additional inputs had been provided by PMB members and would be incorporated in the presentation. Regarding the WLCG Tier-1 Survey results, it was thought that the data had been finessed as much as was possible. The slide was informative, and we had 2-2.5FTE at RAL testing new technologies. It was agreed that it would be good to cover this data as the survey was generally known. PC considered it to be unnecessarily detailed however, and he suggested simplifying the statistics. It was preferred however to provide more detailed information for the Committee. DB would circulate a draft slide within the updated version of the talk. The Alice requirements did muddy the waters. PC suggested omitting Alice, so that it was not counted against the Particle Physics Programme - could we give an analysis excluding Alice? DB thought this was too hard to work out in the context of WLCG. Were there any other comments? The talk slot was 45-minutes, starting at 11.30 am. The PMB members attending would meet beforehand in London.
STANDING ITEMS
==============
SI-0 Monthly Report from Development (Cloud) Group
---------------------------------------------------
There was no report.
SI-1 Dissemination Report
--------------------------
SL reported on behalf of Tom Whyntie:
###Press release on GridPP and the LHC restart
TW will send this out at 12:00pm today. Many thanks to Dave Britton
(DB) and Jeremy Coles (JC) for material and supporting quotes. If anyone would like a copy for their university, or receive grid-related press interest, please contact TW.
SI-2 ATLAS weekly review & plans
---------------------------------
RJ was absent.
SI-3 CMS weekly review & plans
-------------------------------
DC noted nothing major to report.
SI-4 LHCb weekly review & plans
--------------------------------
PC noted nothing major to report.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) In EGI Engage there is effort dedicated for improvements in the Accounting Portal and therefore EGI is seeking to create an Accounting Portal advisory board. Should we be involved?
2) Since last week there have been problems where alarms from the central Nagios are not reaching the regional/site dashboard. Sites have been advised to apply extra vigilance.
3) 12 OpenSSL vulnerabilities were announced last week (https://www.openssl.org). Sites are being advised to patch their systems as soon as possible.
4) EGI is seeking to understand whether any sites have need for supported UMD middleware on SL5 to remain beyond 2015. In the UK we are following up on a few cases where older harder may lead to such a situation - for example for some Tier-2 disk pools.
SI-6 Tier-1 Manager's Report
-----------------------------
AS reported that a hardware delivery was due on Wednesday; he had downgraded the risk in the Risk Register accordingly. Other minor issues were being dealt with.
SI-7 LCG Management Board Report of Issues
-------------------------------------------
PC had attended the MB. An issue had been raised about Data Protection and problems with the USA. Some guidance would be released in due course; details were not available at the moment. This was a good example of the intermittent issues that arose, that GridPP needs effort to respond to. The onus may fall on the Experiments regarding name identity. An updated Data Protection Policy was likely to be produced. At the MB other updates had been provided: relating to WLCG critical services at the Tier-0; on multicore accounting where, in order to get this working sites had to publish the number for non-multicore queues. Sites were aware of this.
REVIEW OF ACTIONS
=================
All actions were ongoing.
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included).
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
ACTIONS AS OF 23.03.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included). Mark Slater at Birmingham had requested a Ganga talk.
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
The next PMB would take place on Monday 30th March at 12:55 pm.
GridPP PMB Meeting 558 (16.03.15)
=================================
Present: Dave Britton (Chair), Andrew Sansum, Dave Colling, Jeremy Coles, Tony Doyle, Pete Gronbech, Tony Cass, Pete Clarke (Minutes Ð Suzanne Scott)
Apologies: Roger Jones, Dave Kelsey, Steve Lloyd, Claire Devereux
1. Oversight Committee (OC) Meeting
====================================
The OC meeting was due to take place on Wed 18th March. Two talks were required: 1) Status Report on GridPP4. PG should give this talk and we should use this opportunity to communication to the OC the issues relating to GridPP5, as there is crossover of personnel on both committees. We must highlight the engagement with H2020 and EGI future plans. The rest was formulaic but certainly non-trivial. We may need updates to reflect GridPP4+.
2) GridPP5 Preparation. DB would give this talk, which would also cover future plans and developments. He was assuming that a 'higher level' view was required. He will raise concern about the evolving process and the parallel supplementary questions which were asked, and which seemed to show a lack of comprehension. Membership of the Review Panel had been denied to GridPP in advance. PC considered that we had an opportunity here to educate the Panel as the questions betrayed a worrying lack of knowledge and understanding of the LHC. We also need the support of the Oversight Committee to rebuff the current approach, which seems to forestall the work of the RRB and the CRSG. TC suggested we should get a statement of support from CERN. This was agreed to be a good idea.
Regarding supplementary question 3, had the larger experiments received this? It was unknown. DB thought we could contact the small experiments directly and provide them with the GridPP5 proposal, to enable them to make an informed decision as to their response? A discussion ensued as to what would work best for the small experiments. It would be impossible for them to guess a percentage cut to GridPP with no background information. PG and DB were preparing the talks for the OC meeting. Six PMB members would attend.
2. GridPP5 Review
==================
This Review meeting would take place on Tuesday 24th March. DB would prepare a talk. Six PMB members would attend.
3. GridPP Posts and FEC
========================
PC had circulated a draft paper. Inputs had been received from DC. PC had tried to define the difference between 'support' and 'researcher' posts. He had also circulated a spreadsheet regarding Grid roles and publications. DB considered that we should show why the Grid posts are similar to RA posts. The FEC is part of the leverage for institutes: there was a standard staff posts/hardware money/resources formula.
4. Wahid Bhimji
================
Wahid was leaving GridPP. The PMB wished to record their thanks to Wahid for all of his contributions to the Project Ð which have been many and varied Ð over many years. The PMB wished him the best of luck for the future.
STANDING ITEMS
==============
SI-0 Monthly Report from Development (Cloud) Group
---------------------------------------------------
DC reported that there had been a meeting last Friday. They have a couple of working clouds at the point where they had now completed most of the proof-of-concept work. The question was, did we want to grow these to a resource that would be large enough for the experiments to use? Or did we want to leave these as a testbed site as now? The experiments were not really calling for resources and there were questions of accounting and manpower which would have to be a policy decision.
DB noted that 90% of Grid resources was dedicated to the LHC experiments. What did they want? What freedom of choice do we have? DC noted that CMS use resources in whichever way they come, they run on the CERN Cloud, but also use Grid resources. They were happy either way. There was no strong push from CMS one way or the other.
DB considered that we needed to discuss this at GridPP34. We should have a discussion session to discuss which was the better solution for the experiments. The VAC solution also worked well. We needed to review options. This could wait until April. The discussion session would need a structure however: small sites with VAC; openstack, etc.
ACTION
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
Regarding the HP Sponsor talk at GridPP34, we had requested that this would be regarding high-density, low-energy, rack-based solutions.
SI-1 Dissemination Report
--------------------------
SL had sent an email report on behalf of Tom Whyntie:
### Articles in "The Conversation" on the grid and the LHC Restart
* TW, with help from Jeremy Coles (JC, who had prepared material for a Nature news article last year), has published a piece for "The Conversation", a nationwide online news source with content provided by researchers. This can be found [here](https://theconversation.com/number-crunching-higgs-boson-meet-the-worlds-largest-distributed-computer-grid-38696).
* Roger Jones (RJ) has also published an [article](https://theconversation.com/upgraded-lhc-pushes-physics-into-the-unknown-38754) about the LHC restart which also mentions the grid.
###CERN@school Eclipse 2015 measurement - data shared on the grid
On Friday 20th March 2015 there will be a solar eclipse in the UK. CERN@school is attempting a nationwide data collection exercise to coincide with this (at the very least, it will give schools something to do if it is cloudy!). BBC Stargazing Live will be covering the eclipse and RAL will be present with a CERN@school detector. CERN@school is making the eclipse data publicly available via the GridPP Collaboration's DIRAC system using http access to SEs. Two test runs have been completed and the data has been uploaded accordingly.
* More information:
http://cernatschool.web.cern.ch/experiments/eclipse-2015
* Eclipse data:
http://hepgrid11.ph.liv.ac.uk/dpm/ph.liv.ac.uk/home/cernatschool.org/daqmap/public/test002/
SI-2 ATLAS weekly review & plans
---------------------------------
RJ was absent.
SI-3 CMS weekly review & plans
-------------------------------
DC noted nothing major to report.
SI-4 LHCb weekly review & plans
--------------------------------
PC noted nothing major to report.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) The topic raised in previous updates, but not discussed by the PMB, about increasing resources behind GridPP cloud instances was discussed at the technical meeting on Friday. The general feeling was that we do not currently have strong reasons to do this, except perhaps if we want increased participation in the EGI Federated Cloud work or wider University central resources to integrate (AM also notes a fork of Vcycle is being written to support OCCI). We have demonstrated working approaches with different cloud technologies and the lightweight VAC approach. The general cloud methods are not easier to deploy and manage than grid middleware but provide useful flexibility. An overview of what is in use is being collated under: https://www.gridpp.ac.uk/wiki/Cloud_%26_VM_status.
2) There was a core-ops meeting last week. The focus areas were documentation - GridPP keydocs (some will be removed and others added); middleware testing (our offer to test machine/job features implementations within WLCG has stalled due to poor/no documentation, and our EGI SR contributions are being updated at https://www.gridpp.ac.uk/wiki/Staged_rollout_emi3); monitoring (the EGI fork is now pursuing ARGO) and wider VOs (the GridPP DIRAC service offering is making progress but behind our December plan).
3) The latest blog postings as of last week across our area blogs (http://planet.gridpp.ac.uk) suggest we might do more in this area:
Storage - 9th March 2015 (Storage accounting) NorthGrid - 24th February 2015 (Condor)
Tier-1 - 22nd January 2015 (CEPH)
SouthGrid - 14th October 2014 (Nagios monitoring non-LHC VOs) LondonGrid - 8th October 2014 (xrootD & ARGUS) ScotGrid - 24th March 2014 (3 coordinators).
4) The draft T2 reliability & availability figures for February have been circulated.
ALICE (http://wlcg-sam.cern.ch/reports/2015/201502/wlcg/WLCG_All_Sites_ALICE_Feb2015.pdf):
All okay.
ATLAS (http://wlcg-sam.cern.ch/reports/2015/201502/wlcg/WLCG_All_Sites_ATLAS_Feb2015.pdf)
QMUL: 68%:73%
UCL: 25%:25%
Sheffield: 88%:88%
RALPP: 80%: 98%
Sussex: 73%:73%
CMS (http://wlcg-sam.cern.ch/reports/2015/201502/wlcg/WLCG_All_Sites_CMS_Feb2015.pdf)
RALPP: 81%: 99%
LHCb (http://wlcg-sam.cern.ch/reports/2015/201502/wlcg/WLCG_All_Sites_LHCB_Feb2015.pdf)
RALPP: 82%:100%
The site problems encountered during February are:
RALPP: Had a scheduled downtime for a dCache upgrade. Note for ops tests (EGI report) there was a problem where a protocol entry for gsidcap was cached in the top BDII for four days after it was removed, causing the GetTURLs test to fail.
Sheffield: Newly installed WNs were online while searching for a configuration problem with them. They were fixed. Site points out that ATLAS ASAP was 95% for February.
Sussex: Resolved the problems mentioned in the January report by 7th February and then got back into Atlas test mode on 8th Feb, and got Multicore queue enabled again on 11th. Site is thought to be stable now.
UCL: No report. (A low availability ticket has also just been raised by our ROD).
5) EGI has polled all NGIs about the sustainability of their CAs. For reference the response sent back for the UK was as follows:
ÒSTFC has operated the UK eScience CA for many years and currently support about 2500 certificates (approx. 2/3 are host certs). We have been financing this internally for the past few years after external funding stopped. Although this support is due to run out at the end of the next financial year (end Mar 2016), we are in the process of putting together a case to continue this funding into the future.
Note that although the UK can obtain Terena/TCS host certificates, we aren't signed up to receive personal user certificates from Terena.
JANET (out NREN) have also started the process of setting up a UK-wide (not just eScience/Grid/Cloud) certificate service (this will not be related to the Terena TCS service) and our expectations are that we will migrate our users to their CA, as and when the time comes. As this will take some time, we have no current plans to terminate the current service and are working with JANET to see whether migration is indeed possible.Ó
6) Wahid Bhimji will be leaving ECDF at the end of this week. Many thanks to Wahid for all his contribution to GridPP ops/storage over the years! A DPM Collaboration meeting is being organized to take place in the coming weeks. Sam Skipsey has already stepped forward to help in several areas.
For information:
A) A summary of the pre-GDB last week is available at: https://twiki.cern.ch/twiki/bin/view/LCG/20150310PreGDB. There was a look at overlaps between eScience areas in the morning and cloud issues in the afternoon.
SI-6 Tier-1 Manager's Report
-----------------------------
AS reported on two key operational issues: 1) hardware delivery still had no firm date; 2) the top of the Tier-1 router crashed in operation, a Review Meeting would be happening soon. A load replacement device had been installed and that had also crashed. The standby router was functioning well.
SI-7 LCG Management Board Report of Issues
-------------------------------------------
An MB would be taking place tomorrow; PC would be attending.
AOB
===
- WLCG Survey
There ensued a discussion on the slides which AS had circulated, as to what the figures meant in each plot. PC asked if some of these slides could now be used? The Tier-1 looked less lean in comparison with other similar sites, however AS cautioned that the data quality was low and unusable. He will ask for a different cut of the data, particularly for the four large sites, but noted that this may not be forthcoming.
REVIEW OF ACTIONS
=================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included).
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
557.1 PG/SL to upload the OC documents to the webpage and verify that the links and passwords work. Done, item closed.
ACTIONS AS OF 16.03.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included).
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
558.1 SL/PG/DC to organise a structured discussion session about clouds at GridPP34; this should be added to the Agenda.
The next PMB would take place on Monday 23rd March at 12:55 pm.
GridPP PMB Meeting 557 (09.03.15)
=================================
Present: Dave Britton (Chair), Andrew Sansum, Dave Colling, Jeremy Coles, Tony Doyle, Steve Lloyd, Dave Kelsey, Claire Devereux, Pete Gronbech, Roger Jones (Minutes Ð Suzanne Scott)
Apologies: Tony Cass, Pete Clarke
1. Oversight Committee (OC) Documents
======================================
PG hoped to submit these today. The OC would take place on 18th March. The GridPP5 Proposal would be reviewed at the meeting on 24th March. PG had circulated the most recent version of the OC Documents: the Financial Status document followed the same format as previously. He had updated the tables. He had inserted estimated costs for GridPP5. 10 out of 13 sites had already submitted their JeS forms today. The Project Status document was also completed following various iterations. We were now ready to submit these documents. PG/SL to upload them to the OC webpage and verify that the links and passwords work; then we would submit them this afternoon.
ACTION
557.1 PG/SL to upload the OC documents to the webpage and verify that the links and passwords work.
2. STFC Supplementary Questions
================================
Two additional questions had been asked by STFC. Re Question 1, this was on AS, and he had circulated a draft response. Those who framed the questions seemed to be looking for a simple algorithmic cost for the Tier-1 which was not appropriate or feasible. No simplistic model was possible but we did have a suite of complex and detailed spreadsheets. DB would take AS's response and combine it with his own reply to Question 2. On the staff side, we did have spreadsheets which showed how the staff effort supports the service which GridPP provides. DB would draft a response.
On Question 2, they seemed to want a figure that represented the UK Computing commitment per RA. Again, this question was neither appropriate nor feasible. It would be best put to STFC who would need to decide what was 'in scope', however it would give a meaningless number in any case which could not be used for anything. DB would respond to the supplementary questions.
3. JeS Forms
=============
PG advised that the deadline for submission was 10th March. Very short notice had been given by STFC for the production of these figures, and it was recognised that the short timescale was challenging for local institute admin. TD reminded that late submissions would not be allowed on the electronic system. PG would circulate a reminder email. The hardware was pending until after the GridPP5 Review.
4. GridPP5 Proposal
====================
DB noted that he was currently working on Version 17. SL had helped with document length over the weekend. Any outstanding actions on members re this document were required immediately. DB considered that all components were extant at this stage. DB was currently addressing the margin comments. It was agreed to add a paragraph on 'contingency' and 'working allowance' and the approach to these was agreed. For the Financial Summary, tables had been added. There were space constraints this time round with a maximum page limit. DB would generate urgent actions from his next read-through of the current document.
STANDING ITEMS
==============
Due to the prevailing pressure of producing the OC Documents; the GridPP5 Proposal; and the Institute JeS forms, it was not possible to review the Standing Items.
SI-0 Monthly Report from Development (Cloud) Group
---------------------------------------------------
There was no report.
SI-1 Dissemination Report
--------------------------
There was no report.
SI-2 ATLAS weekly review & plans
---------------------------------
There was no report.
SI-3 CMS weekly review & plans
-------------------------------
There was no report.
SI-4 LHCb weekly review & plans
--------------------------------
PC was absent.
SI-5 Production Manager's Report
---------------------------------
JC reported as follows:
1) As per my message circulated on Friday [27th February], GridPP cloud instances are ready to receive additional resources (providing the experiment accounting is reconfirmed as accurate) and a strategy decision is needed at this point to put more resource behind the cloud instances or not. The direction of the GridPP cloud resources at Imperial also needs discussion.
2) As part of a more formal rollout of cloud/VM resources and services Andrew McNab has suggested that we incorporate a new core-ops task in this area. We can start this activity with GridPP4+.
3) VOMS-admin has replaced VOMRS at CERN. All the experiments are encountering minor problems which are being addressed.
4) ATLAS has indicated that with MC15 starting the experiment considers Run2 to have started for the computing.
5) We encountered dashboard problems (regional and security) due to GOCDB updates not being reflected in the CIC Portal by the time of the change.
6) Sites were slow to pick up on a release of CA distribution version 1.62 (23rd February) as the announcement fell victim to a trial of reducing EGI broadcasts. The problem was only noticed when sites started failing CA tests last week.
For information:
A) There is a pre-GDB this week with a focus on HEP and other sciences in the morning and cloud issues in the afternoon: https://indico.cern.ch/event/319819/.
B) The agenda for WednesdayÕs Amsterdam GDB: https://indico.cern.ch/event/319745/timetable/#20150311.detailed includes talks on future EGI/WLCG work, European procurements, WLCG operational costs and supporting other sciences.
SI-6 Tier-1 Manager's Report
-----------------------------
AS noted nothing major to report.
SI-7 LCG Management Board Report of Issues
-------------------------------------------
There was no MB.
REVIEW OF ACTIONS
=================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included).
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
ACTIONS AS OF 09.03.15
======================
533.3 DK to draft a preliminary response for any future security incidents where the press might contact GridPP (Tom Whyntie should be briefed on this as well in case he is contacted direct).
549.2 AS to update all sites who hold Extreme kit, regarding the outages over the Christmas period.
550.7 SL/PG to consider the Agenda for GridPP34 (theme is Run-2; 'other users' non-LHC talks to be included).
554.1 RJ/JC to take forward the Sussex situation regarding their Tier-status.
554.4 DK to contact EDFA JET and discuss a possible move to a more lightweight model than a Tier-2 site.
557.1 PG/SL to upload the OC documents to the webpage and verify that the links and passwords work.
The next PMB would take place on Monday 16th March @ 12.55 pm.
|
