NT1/SG-Meeting-2019-11-13

From neicext
Jump to navigation Jump to search


NT1 meeting 2019-11-13, 09:30 NordForsk, Oslo

Invited: Mattias Wadenstein, HPC2N; Oxana Smirnova, HEP, LU; Michaela Barth, PDC-HPC

Location: https://umu.zoom.us/j/232177985

Agenda

  • Presence and approval of agenda
  • Report highlights
  • Staff changes
  • CA status
    • In relation: Update of ToRs
  • NORDUnet network contract
  • WLCG workshop summer 2020
  • NT1 report to NeIC board
  • Tracking of production start times for new hardware
  • Open actionpoints
  • AOB

Minutes (Approved)

Live Minutes: https://docs.google.com/document/d/1-FwZAcC-s9Q9YuM8I62aJPlflfhP_sA57ZC5wCGtN2E/edit


Presence and approval of agenda

Presence: all.

Decision: Agenda is approved.

Report highlights

https://indico.neic.no/event/99/contributions/354/attachments/147/237/20191113-NeIC-NT1-2019Q3-Status-Report.pdf

New ALICE disk in Bergen finally in production: We are yellow. Traffic on the new network in Norway can now be seen.

Tape: Bergen wants to decommission the old tape before taking the new one in production. In between take the storage on disk for a “few” months. Timeline for tape is worrying, but only 2.2 PB. Not urgent for ALICE, but still. This has to be operational when the shutdown is over (expected Nov. 2021 according to current public schedule).

Lots of time spent investigating performance issues leading to slower commissioning of ALICE CPU in Bergen

Yellow because of Norway (both for tape and CPU): ALICE installed tape < 90% of pledge–ALICE used CPU < 90% of pledge (Finland’s contribution is too small to be relevant)

Pledges needs to be fulfilled. During shutdown no one is getting hurt, so it is hard to push. This should be discussed with the NLCG.

Shutdown could be seen as a golden opportunity to test and setup new resources.

Staff changes

24/7 contracts getting renewed, already signed by NordForsk, staff signing starting this week.

Upcoming contracts running out: Vincent Garonne in February. Currently on 100%

New operations recruitment: 2 applicants to schedule interviews for. In person interview preferred. Locations close to candidates: Michaela to host an interview in Stockholm, asking Gudmund for hosting an interview in Oslo.

  • AP Mattias: schedule interviews with applicants.

CA status

Ready to sign: DK, NO, NordForsk with suggested change in ToR.

No Feedback from FI.

Late Feedback from SE that they can’t sign in the current form, input promised earliest for the end of this week.

  • AP Mattias to contact SE and offer his help to go through the CA document.
  • AP Michaela to inform CA signees about status.

In relation: Update of ToRs

https://docs.google.com/document/d/1VlbhuAxVByvQ2GGFue7PmATzW7JjblUDBVhB8ALgWl8/edit

Decision: Suggested new wording is approved.

NORDUnet network contract

https://wiki.neic.no/int/LHCOPN_Contract

Slovenia added, presentation item changed/removed. Network and performance monitoring added. Name change of NordForsk director needed.

For the previous contract we did an external evaluation showing that the costs were justified, we can rely on that and compare to that. Géant cost reduction was forwarded to us. Network endpoint performance monitoring currently a lump sum, since we are piloting this. This contract is cheaper than the previous one.

How much is our share in the NORDUnet budget? A comment on the actual amount of FTEs and percentage is desired as comment to NorduNet project participation and hosting cost.

Actual routers handle not only LHCOPN network, thereof virtual routers.

  • AP Maswan: Find evaluation and link from wikipage for reference.

Decision: Besides suggested changes the draft is in a good enough shape to be shown to NordForsk for signing.

WLCG workshop summer 2020

Hosted in Lund in May, no facilities available in Umeå. Other conferences at the same time. Scandic hotel outside of Lund. Programme arranged by international PC

LOC: Caterina Doglioni, Oxana, Mattias

International workshop

NeIC visibility can be reached in other ways than with sponsorship. Commercial sponsors preferred. Could worktime count as contribution for sponsorship? _Keynote_ or welcome talk on Nordic collaboration could be a good idea.

  • AP Oxana, Mattias: Bring up discussed ideas with WLCG workshop LOC.

NT1 report to NeIC board

Now scheduled for spring.

Document started at: https://docs.google.com/document/d/1qb9UdtnB5lPDBkBKnX4-lNd1Y0Q6_Mn8_efK0_TcYQ8/edit

Template with link to example: https://docs.google.com/document/d/1eQXNBQHrYgWUD0e-LvLRiBzmmu7uKrYLOq43EVKkh9Q/edit Another example will be the iOBS midway report https://docs.google.com/document/d/1MjMrprbsdSiAceSgPhVFNEjEeZ7bVVQc4CgoZBZL7fw/edit


Mattias having time first week of December to think about content and agree on a draft. Next steps should include computing model for HL-LHC.

  • AP Michaela to clean up her first thoughts until end of November
  • AP Mattias to sketch NT1-update content still in December

Oxana to read through it then in January. Ideally document gets finished end of January.

Onboarding new communities document: https://docs.google.com/document/d/1iyTaHZGua3hkjd_2YYxrYOKF8nn6BzmH6BenGPkh6xw/edit

30-50 kNOK reserved in budget for power and cooling

  • AP Oxana to go through onboarding document still this year.


Tracking of production start times for new hardware

Make a wikipage where we track commissioning time for new hardware and for how long they plan to be in service, particularly dCache pools. So far we relied on sites to run this professionally but this needs more tracking. We have one website for central resources, nothing for outlying resources.

  • AP Mattias to create Tier-1 pledged hardware tracking website

Open action points

New:

old:

  • AP: Mattias to prepare a basic list on who is doing what within NT1 context. (done)
  • AP: Oxana to write down an advertisement pamphlet on what the NT1 could offer to those communities so they end with a nice coherent service offering that fits into the international collaboration, in case they will seek support for a distributed e-Infrastructure (and apply for funding)-> (in progress: Oxana created a first slide; distributed to SG via email)
    • Distributed Nordic operations team
    • Consultancy on how to design distributed storage
    • Advice on how to set up international operations team
    • Consultancy on how to arrange network and procure those services (we can not overload our own service agreements with WLCG and NORDUnet)
    • Add comparison with other known offers
    • next step: Mattias to provide more input
  • AP: Mattias to make operational procedures again public and up to date so it can be referred to. (in progress)

AOB

  • Oxana to inform of changed NordForsk entrance to NLCG attendants.
  • Mattias to officially advertise AHM, Michaela already announced it in chat and last Friday’s operational meeting.
  • Impressions CHEP: Maiken’s talk was well received and even mentioned in concluding summary. Discussion on how to do datacha caching and improving throughput.

xCache (American) which is just a proxy not actual caching was pushed very hard. Kubernetes was everywhere. DataLakes not as hyped any more. Discussions on ALICE completely switching to new software model.