Choose timezone
Your profile timezone:
IT-protoDUNE coordination (Horizonal and Vertical Drift)
→
Europe/Zurich
Description
Chair: Jose Castro Leon Secretary: Elena Gazzarrini
Regular meeting between IT department and computing experts from protoDUNE. The meeting goals are:
- Exchange technical information related to computing for the Experiment
- Provide requested technical details on IT services
- Highlight problems related to IT services
- Follow up on items raised at the meetings and record them on an action list
Coordination: JIRA, Minutes, etc. https://twiki.cern.ch/twiki/bin/viewauth/ProDUNEIT/WebHome
Zoom meeting room:
https://cern.zoom.us/j/95006476108?pwd=NjIyVGRtZy92SUROaVVPeGNQcXRtdz09
Meeting ID: 950 0647 6108
Passcode: 08430992
Videoconference
IT-protoDUNE coordination (Single Phase and Double Phase)
- Zoom Meeting ID
- 95006476108
- Host
- Ignacio Coterillo Coz
- Alternative host
- Pablo Saiz
- Useful links
- Join via phone
- Zoom URL
-
IT-protoDUNE coordination meeting 07-09-2022
Attendance
- Steven TIMM (protoDUNE)
- Denis PUGNERE (protoDUNE)
- Elisabetta PENNACCHIO (protoDUNE)
- Xavier ESPINAL (IT)
- Jose CASTRO LEON (IT)
Coordination News
RCS-IT Engagement
- (Xavier) Explain a little bit the Engagement body
- (Xavier) We would like to explore the possibility to join the Small and Medium Experiments meeting as we see that you could benefit from the experience of other experiments
- (Xavier) We can invite to the next couple of meetings on SME and then evaluate later if you see that it fits you. We'll keep this meeting in the mean time
- (Jose) I'll invite you to the next one in October
Cloud Network Migration
(Jose): The cloud team has just announced the schedule for the VM cold migration campaign during today’s session on ASDF. The cloud team foresees to do the bulk part of the migration between September 15th and October 31st. Users of Cloud VMs affected will receive their scheduled migration times by mail.
(Jose) For service VMs, they will also receive a link to allow them to propose alternative migration dates to avoid interruption on their respective services.
(Jose) For personal VMs, the migration schedule will be fixed. As a reminder personal projects are intented for personal and/or development purposes.
(Jose) In the rare circumstance that an user group/experiment has located their important VM on a personal project, we encourage to open a ticket to the Cloud Infrastructure using the normal channels (serviceNow) so we can propose a proper solution for each special case.
(Jose) List of machines affected:
| VM Name | Landb Owner | LanDB Responsible | Project | Experiment | Migration Date | Flavor | Estimated downtime |
|---|---|---|---|---|---|---|---|
| np04-webgw1 | dune-comp-vm@cern.ch | dune-comp-vm@cern.ch | DUNE | DUNE | 2022.10.05 0:00 | m2.large | 30 min to 2 hours |
| p3s-web-new | dune-comp-vm@cern.ch | dune-comp-vm@cern.ch | DUNE | DUNE | 2022.10.10 0:00 | m2.large | 30 min to 2 hours |
| p3s-web | dune-comp-vm@cern.ch | dune-comp-vm@cern.ch | DUNE | DUNE | 2022.10.14 8:00 | m2.medium | less than an hour |
| dune-vm-build-02 | dune-comp-vm@cern.ch | dune-comp-users@cern.ch | DUNE | DUNE | 2022.10.17 0:00 | m2.large | 30 min to 2 hours |
| dune-vm-build-03 | dune-comp-vm@cern.ch | dune-comp-users@cern.ch | DUNE | DUNE | 2022.10.20 8:00 | m2.xlarge | 2 to 4 hours |
| p3s-db | dune-comp-vm@cern.ch | dune-comp-vm@cern.ch | DUNE | DUNE | 2022.10.21 0:00 | m2.medium | less than an hour |
| p3s-content | dune-comp-vm@cern.ch | dune-comp-vm@cern.ch | DUNE | DUNE | 2022.10.28 0:00 | m2.medium | less than an hour |
If you are OK with those dates, please approve
https://gitlab.cern.ch/cloud-infrastructure-migration-campaign/dune/-/merge_requests/1
Atlasian change of license policy (JIRA, Confluence)
- (Jose) Explained a little bit about the issue and its possible impact
- They are not affected as they are not using any of those tools
IT Information Update
Round table and action list review
- (Steven) Horizontal drift is a very bit behind, it will start collecting data
- (Steven) We’ll not see beam until the new year
- (Jose) There are some questions to pass around for GPU usage?
- (Jose) Like if you are using CERN GPU resources and if that’s the case, if you are happy with those
- (Jose) The other question we have is a quantification of their expected needs for the next years from CERN IT
- (Steven) new detector will need ML
- (Steven) tricky to find a way to use it in the current framework
- (Steven) Some more in batch, we have a small part in GPUs, nice to have more
- (Steven) some resources for training, some for inferencing
- (Steven) big places have vGPUs (AMD, intel)
- (Steven) Do you have the pointer to the future architecture meeting (ask Xavi for the pointer)
- (Steven) We need someone to do it and provide input
- (Steven) Yesterday, we started to do a big data challenge (generated 500tb of simulated data)
- (Steven) Testing workflow allocator (running with data that is closest to it)
- (Steven) Reading a bit more at this moment than usual
- (Denis) Not for me
- (Elisabetta) For vertical drift, we are resuming data taking end of the month, and it will be one month of test
- (Elisabetta) so at this moment there is no data coming onto EOS
- (Elisabetta) concercing mail from Kirby, he is not on some of the egroups
- (Steven) we need to add him to that one
- (Elisabetta) or Kirby need to add himself to those groups
- (Xavi) Jose please follow with Luca about the egroup memberships
- (Jose) Yes
Current Actions
- None
New Actions
- Check permissions with EOS team in protodune space (Jose)
AOBs:
- Next meeting called for:
- October 5rd at 3pm hopefully hybrid
There are minutes attached to this event.
Show them.
