GriddPP PMB Meeting 824

Written by

GriddPP PMB Meeting 824

Present: David Britton (Chair),Peter Clarke
David Colling. Davide Costanzo, Alastair Dewhurst, Tony Doyle, Katy Ellis, David Kelsey, Peter Gronbech, Jonathan Hays, Roger Jones, Steve Lloyd, Andrew Sansum, Sam Skipsey (Minutes), Jill Sambrook (Minutes)

Apologies: Tony Cass, Andrew McNab

ITEMS

1) GridPP6 OsC meeting- feedback and discussion [DB]

GridPP6 Oversight Committee (OsC) Feedback and Actions document was circulated by DB on Wednesday 6th July. The group then discussed the actions below:

Action 1: Clarify the timetable for GridPP7 with the Collaboration. The group aim to focus on GridPP7 plans at the Ambleside meeting next month.

Action 2: Collaboration to provide a short document (a few pages) outlining the vision for GridPP7 by the end of July.
This was done by the end of June and was marked as complete by the committee.

Action3: Collaboration to provide brief, narrative information on data recovery processes in the report for the next OsC meeting.
Action for SS to complete. Sam has already been working on document and so has gathered quite a lot of information. DB said it would be good to look at this from an experiment perspective. CMS statement could compliment this and ATLAS information could also be good.

Action 4: Collaboration to explore GridPP taking part in IRISCAST and if possible, provide a status report for the next OsC meeting.
JH stated that Queen Mary is officially funded on the project and this therefore covers GridPP’s participation. It was agreed by DC that Imperial would also like to take part in IRISCAST. DB agreed this would be a good as Imperial is the one major site that is significantly different than the other main sites.
Thanks to both.

Action 5: Collaboration to provide narrative in the OsC report on the process that is used to determine the networking costs for GridPP6 for the next OsC meeting
Action moving forward was for DK and AD to draft a short paragraph providing this narrative. DB asked for there to be a brief statement for GridPP6 with more details and plans for GridPP7 to be highlighted.
DK and AD can send draft to DB.

Action 6: Collaboration to request a user report from the ALICE experiment for all future OsC reports.
Some discussion around who was best to approach to provide this report.
It was agreed, for now, DB would speak to someone at Birmingham and explain the request. DB can provide an example of what is needed and if they would be able to provide this and then DB could Liaise with AD.

The next OsC meeting will be held in November 2022. We will likely have to submit documents by mid-October and so the group can start focussing on the next report around Mid-September.

The feedback was really good from the committee with a lot less actions than the previous review, so thanks to everyone for all of the work done.

2) GridPP48 Ambleside [SS]

47 registrations so far, so we are not too far off our usual figures for these events.

No update on accommodation. Still no response from the venue. Roger to forward the last correspondence with the venue to Jill and Jill/Sam to now try and organise this.

Sam has now made some progress with the agenda and speaking with various groups including LZ and SKA.
The networking session is probably going to take up a big chunk of the Thursday morning.
Sam and Jill to do some more of the detailed planning on this over this coming weeks.
DB suggested Jill liaise on the phone with the venue (more responsive) and document detail/agreements in email.

Alastair is still working on timing of requests for sponsorships.
DB suggested the group add a page about sponsorship. “Call for Sponsors”, to the GridPP website. This would highlight information about sponsorship and how to go about this. It would mean PMB members can refer companies to this in the future and might be helpful moving forward.
DB and SS to draft content and circulate to the group for approval before this is added to the website.

3) AOCB:
PC informed the group that Chris Walker had now joined JISC and would be attending the Ambleside event.

STANDING ITEMS

SI-0 Tier-1 Manager’s weekly report & Technical Meetings [AD]
• Technical meetings

  • SS and Duncan R still happy to chair Perfsonar meeting and have this is in hand.
  • AD to ask Tim Noble to chair Rucio meeting as DC has been unable to arrange this.
  • AD to look into New OS meeting. . Perhaps something to discuss at GridPP48 – add this to the agenda. Perhaps include this in tier 1 presentation

• Operations update
Things have been fairly stable in the operational side of things.

  • LHCb “slow stats” issue – stats via xrootd protocol 5x slower than stats via davs (for all sites using xrootd servers for both). Jyothish is following up with Chris H for LHCb resolution.
  • “failed Ops tests” – Ops tests include directory listing test (which does not make sense for an Object Store). RAL discussing approach to removing this test for them.
  • FTS3.12 update coming
    • Update on longer running issues.
    Brief updates provided.
    Slow Deletions – average deletion is not slow, but there are outliers (and tails are long). ATLAS, for ex, sees less slow deletions because they have shorter timeouts (so they see either “fast enough” deletions or timeouts which retry), LHCb with longer timeouts sees slow deletions in the tail. Work actively underway in trying to understand the causes of the long tails in deletion time so it can be ameliorated.
    Vector Reads – following a few avenues. Rob was able to replicate the problem. Will arrange another meeting with Rob shortly and keep working on issues.

• Contingency planning in place for upcoming high temperatures. Worse case we might have to reduce capacity for a short term. Will try and avoid this.

• Procurement has been somewhat challenging.
AD has now placed an order for storage capacity and CPU.

SI-1 ATLAS Weekly Review and Plans [DCos]
No major issues to report.
DCos and JH mentioned that QMUL are still reduced to 10GB due to networking problems – for a brief period this was much lower due to config issues on the limit set but this is now fixed.. This will remain capped at 10G until September. Should be able to go to 100G.

SI-2 CMS Weekly Review and Plans [DC/KE]
Operations going fairly smoothly.
Still seeing occasional long read time on some runs. Will continue to investigate this.
13.6TB data coming for archive.
Deleted 5TB from tape to make a bit more space
Continuing to test Xrootd-monitoring-tool Shoveler. CERN machine had to be restarted so still working with this.

SI-3 LHCb Weekly Review and Plans [AM]
Alistair covered this (Rob update) PC.

SI-4 Operations Meeting Report [SS,PG,PC] –
NTR that hasn’t already been spoken about above.

SI-5 LCG Management Board Report of Issues [DB]
WLCG MB delayed for 2 weeks

SI-6 External Contexts (eg NGI/EGI) [PC/JH]
NTR.

REVIEW OF ACTIONS
800.5 AD – Arrange in person DIRAC/Rucio meeting at IC (Jan22). [on-going] – Plans to re-establish virtual meetings and then set up an in person meeting. AD to ask Tim Noble if he can arrange this.

818.5 – DB/SS Group to formalise VOs to be added to the approved list. Ongoing
DB and SS have begun to look into this. 3 in particular were flagged as perhaps needing to be removed.
NOne
MaGrid (Morocco institute)
ComplexSystems
Raised with TBSupport for site feedback on who supports these and SS will wait until he has some feedback about use.
Discussion also around Microboone.

822.1 – AD to provide update about sponsorship for GridPP48 Ambleside.
No progress, but will update. DB and SS working on web content. See action 824.5 below.

823.1 – AD to Invite Rob to the Wednesday Liaison meeting for a few weeks to discuss LHCb tickets and leave as a standing item at the start of the agenda for a few weeks.
Will do this now.

824.1 – SS will take lead on OsC action:
Collaboration to provide brief, narrative information on data recovery processes in the report for the next OsC meeting.

824.2 – DK and AD to draft a short narrative addressing OsC action:
Collaboration to provide narrative in the OsC report on the process that is used to determine the networking costs for GridPP6 for the next OsC meeting
DB asked for there to be a brief statement for GridPP6 with more details and plan for GridPP7.
DK and AD can send draft to DB.

824.3 – DB will take lead on OsC action:
Collaboration to request a user report from the ALICE experiment for all future OsC reports.
Some discussion around who was best to approach to provide this report.
DB to contact someone at Birmingham and start the conversation about this. DB can then liaise with AD moving forward.

824.4 – JS and SS to now contact organisers in Ambleside to make practical arrangements, including trying to sort accommodation queries.

824.5 – DB and SS to create content calling for sponsors to be added to the website. Content will be circulated to the group for feedback before this is added to the website.