BiLD-Dev
Description
Bi-Weekly "Loyal" DIRAC developers meeting. And, following, the LHCbDIRAC developers meeting.
Zoom: BiLD
https://cern.zoom.us/j/62504856418?pwd=TU1kb01SOFFpSDBJeWVBdU9qemVXQT09
Meeting ID: 62504856418
Passcode: 12345678
BiLD – 14/12/2023
At CERN: Andrè, Alexandre, Federico, Christophe, Cedric
On Zoom: Christopher, Janusz, Daniela, Simon, Andrei, Ewoud, Hideki, Lorenzo, Janusz
Apologies:
Follow-up from previous meetings
- Last BiLD 3 weeks ago
- Trying to catch all updates below
- Last DIRAC certification hackathon 2 weeks ago
- Some progress but still issues with jobs
- Federico I did not have time to do further checks this last week
- BildX on Dec 7th
- NTR
DIRAC communities roundtable
LHCb:
Federico+Alexandre+Alexey
- Running on latest v8 patch – NTR
Belle2
Cedric, Hideki
- Pilot
develversion tried and successfully - Trying latest DIRAC v8 with latest DIRACOS2.
voms-proxy-initfails at voms server at KEK (using SHA1). Not failing at DESY.- Chris B prob not using 2048 bits. Please open issue or discussion with details
ILC/Calice/FCC
Andre, Lorenzo
- Using opensearch dashboards from the repository
- Fixed stored accounting/monitoring plots using the fixed script from the v8.0 migration notes
EGI
Andrei
- updated HTCondor utils package to 9.0.19. Requirement for legacy VOs.
- Chris B will make a new DIRACOS release, should pick up the latest version automatically
GridPP:
Simon, Janusz, Daniela
- Nothing to report. We are slowing increasing our token coverage.
Topics from GitHub/Discussions
only un-answered topics with discussion updates:
- Queues with different OS versions
- solution already exists – answer chosen. Further comments?
- Proxy renewal of user jobs
- Federico answered, and asked question – no further replies
DIRAC releases
- v8r0
- v8.0.33
- WorkloadManagementSystem
- FIX: (#7359) added SiteDirector option for CVMFS_locations
- FIX: (#7302) JobAgents will set jobStatus=Failed/Payload failed if and only if the job was previously Running
- Core
- NEW: (#7342) Introduce DIRAC_MYSQL_CONNECTION_GRACE_TIME to specify the grace time of the MySQL connection pool
- TS
- NEW: (#7336) add an index on Status and Type for the TransformationDB.Transformations table
- WorkloadManagementSystem
- v8.0.33
- v9r0
- Created several tags in the last/releases in the last few days, then manually removed part of them. The purpose was to test that the upload to CVMFS could be done from Github Actions (still failing to do so)
DIRAC projects
DIRAC:
Issues by milestone:
-
v8.0:
- 15+ open issues
- usual reminder for closing/moving old ones ( Federico I did close a few yesterday)
- 15+ open issues
-
v9.0:
- 15+ open issues
- Those for Elastic Parameters DB are under
- from previous meeting https://github.com/DIRACGrid/DIRAC/issues/7260
- Chris will take care
PRs discussed:
*
WebApp:
- NTR
Pilot:
- Sadly, no merge of
develtomasteryet. At this point it will be done by next year. - Belle2 said they tried the
develversion and they are happy with it - Will first test in LHCbDIRAC certification (today/tomorrow or next week) then LHCb production setup
- Federico added some documentation too
DIRACOS:
- New rel probably today. For HTCondor and some Globus issue (for which CB created a patch upstream).
Documentation:
- Some random failures – did not understand what happened.
OAuth2:
- NTR
management
- from previous meeting 3 issues left, still valid?
- Always upload releases to CVMFS is again valid for clients
- higher priority now. This should also be done for non-x86 versions.
- Federico Tried but did not manage yet. Some not-understood (gsi)ssh error
- Also the upload of pilot files to CVMFS could be done same way
- Right now the cron-job runs hourly
diraccfg
- NTR
DB12
Alexandre
- from previous meeting
- Ewoud opened PR for py3.11 (in progress)
- missing the Intel vs AMD
- Federico Can we have DB12 run on ARM
- Andrè : I access two “real” ARM machines through OpenStack that can be used for the tests
- Federico: I will pass it to Ewoud
- Ewoud opened PR for py3.11 (in progress)
- We need to understand exactly which factor to apply:
- Alexandre applied a py3.9 factor for being “equal” to py2.7
- this ATM we are not using because we moved to py3.11
- in LHCb we have a “DB16” factor on top of that for being “equal” to HEPSpec06
- Ewoud was applying a new factor to be “equal” to HEPScore
- Alexandre will open a Discusssion in Github to understand if other VOs would be fazed by DB12 score changes
- if not, we can probably strive at making again DB12 score a cheap replacement for HEPScore
- ARM should be also added
- Alexandre applied a py3.9 factor for being “equal” to py2.7
Rucio
- Cedric I have a PR to push for the MetaData catalog
Tests
- Federico I will create a new repo for running “nightly” system tests targeting the DIRAC cert setup
Release planning, tests and certification
-
Certification machines
- lbcertifdirac70 machine:
- we should move to Alma9 box in Q1 2024 (CC7 EOL June 2024)
- As said, this should be done outside of CERN
- we should move to Alma9 box in Q1 2024 (CC7 EOL June 2024)
- lbcertifdirac70 machine:
-
Next hackathon(s)
- next year!
AOB
Next hackathon: Jan 18th
Next BiLD: Jan 25th
Next BiLDX:
-
DIRAC+X hackathon:
- https://indico.cern.ch/event/1343151/
- 23rd-24th Jan 2024
- CERN, “usual” room
- You need to register
- Anyone’s welcome to suggest specific topics
- Federico attempts to upgrade to v9 pre-versions for your certifications setups can be done
-
Next DIRAC Users’ workshop in Lyon https://indico.cern.ch/event/1341205 June 19th
LHCbDIRAC
- v11.0: deploy board in https://trello.com/b/Ep0PAkbv/deploy-110
- NTR
- LHCbPilot to CVMFS
- Federico MR open, merged. Working OK.
- Bookkeping:
- few issues accumulating
- Alexey looking into them
- Downtime of the Bookkeeping:
- we can shutdown the BKK components, and FWDISET requests will be created
- VOMS backend will be shutdown in June 24
- we need to make sure that all links will be updated
- Christophe will do adiabatically over the coming weeks
- OpenSearch migration
- Chris can we drop some indices to ease the migration?
- older logs can be removed?
- Chris can we drop some indices to ease the migration?
- Federico Alma9
- CC7 EOL June 24
- lxplus alias switched to EL9 one week ago: https://cern.service-now.com/service-portal?id=outage&n=OTG0076556
- We probably need to update LHCbDIRAC servers – not sure we can manage to avoid it!
- certification
- RPMs
- CernVM5?
There are minutes attached to this event.
Show them.