BiLD-Dev
→
Europe/Zurich
Description
Bi-Weekly "Loyal" DIRAC developers meeting.
And, following, the LHCbDIRAC developers meeting.
BiLD (Bi-weekly DIRAC Development meeting) – 11/06/2020
At CERN: Nobody, of course!
On Vidyo: Federico, Andrii, André, Christopher, Christophe, Igor, Daniela, Hideki, Janusz, Alexandre, Simon, Cedric, Aymane, Marko, Andrei, Vladimir, Ueda, Xiaomei
Apologies:
Follow-up from previous meeting
- Pilot3 and user jobs needing to run dirac-commands:
- an operational solution proposed on the forum (https://groups.google.com/d/msg/diracgrid-forum/tcDu1glXJcQ/hIBZbSX7AwAJ)
- not working fully, see https://groups.google.com/forum/#!topic/diracgrid-forum/kPcMb1ZTcS0 so PR https://github.com/DIRACGrid/Pilot/pull/108 was proposed and merged
- and failed!
- new PR(s) ready for being merged, this time look better merged
DIRAC communities roundtable
GridPP:
- Few cloud fixes.
- Made some progress on HTCondorCEs (after some suggestions from André)
- problem on both sides (DIRAC and HTCondorCEs)
- full leghth of the proxy to get: under
Resources -> Computing -> HTCondorCE
had to addExtraSubmitString = delegate_job_GSI_credentials_lifetime = 0
- full leghth of the proxy to get: under
- problem on both sides (DIRAC and HTCondorCEs)
CLIC:
- Migrating services to new servers
- Centos7, diracos, still v6r22
LHCb:
- Been using more and more M2Crypto, few bugs here and there
France Grilles:
- Smoothly running
EGI:
- NTR
Belle2:
- Migration to v6r22 should be done next week
Nica:
- Tested first part of https://groups.google.com/forum/#!topic/diracgrid-forum/tcDu1glXJcQ
- will test the DISABLE_WATCHDOG_CPU_WALLCLOCK_CHECK env variable
- Asked if it would be possible to simply do “source bashrc” for “installing/setting up” DIRAC on CVMFS
- For vanilla DIRAC it could be possible, but then you need to anyway supply a specific configuration for the VO
- Something to be added in management tasks?
- For vanilla DIRAC it could be possible, but then you need to anyway supply a specific configuration for the VO
Juno:
- NTR
Current situation
DIRAC
- v6r22:
- NTR
- v7r0:
- created v7r0p26
- Fixes for Watchdog CPU evaluation
- Added a new DISABLE_WATCHDOG_CPU_WALLCLOCK_CHECK env variable that can be passed in (for sloowww sites)
- created v7r0p26
- v7r1:
- v7r1p4 created
- v7r2:
*
WebApp:
- New responsible for the WebApp as Zoltan has left: Andrii
- Looked at all the issues
- Tests fixed for v4r0 and v4r1
Pilot3:
- etc/dirac.cfg linked to pilot.cfg if the etc/dirac.cfg does not already exist (see PR https://github.com/DIRACGrid/Pilot/pull/111/)
DIRACOS:
- xroot5?
- Chris looking at it, incompatibilities with gfal. A release candidate has been created
- Chris B: the latest version of Gfal is broken, waiting for a fix
- Andrei: is diraccfg included in DIRACOS?
- A: it’s in v1r12
VM:
- Should use PoolCE instead of running multiple pilots
Documentation:
- Chris: in LHCb pip does not do a proper dependencies resolution (the six version was not up-to-date). Sorted out by specifying a specific version.
OAuth2:
- Being tested in EGI framework
tornado and other externals
- NTR
management
- https://github.com/DIRACGrid/DIRAC/issues/4604
- in general: we should be doing more automated deploys
- also asked to put pilot files on CVMFS
- see Nica part
diraccfg
- NTR
Release planning, tests and certification
- Not much done wrt to what was discussed 2 BiLD meetings ago (which is still valid).
- PyTests failing for PoolCE: suggested to remove the coverage computation (Andrei)
- https://github.com/DIRACGrid/DIRAC/issues/4631 (Multi-VO Testing) : created task for moving the certification server to multi-VO
- will be done anyway after we have run the first certification hackathon
- André: do we want to have a second setup for multi-VO?
- Federico: we’ll have only 1 setup
Weekly development(s) focus
NTR
DIRAC: current PRs and tasks being worked on, or topics from Google forum
PRs:
- v7r0:
- NTR
- v7r1:
- BDII2CSAgent and non-SRM SEs: not much to be done if Glue1/2 does not publish correctly
- v7r2:
- Nothing that was not already discussed before
On issues:
- https://github.com/DIRACGrid/DIRAC/issues/4616 (Set XRD_RUNFORKHANDLER in RequestExecutingAgent and remove from the bashrc) : waiting for xroot developers first (not connected with the other xroot5 issues mentioned in DIRACOS part above)
Topics from Google forum:
- CLIC: the Framework/MonitoringDB got corrupted, not sure when nor why. At the same time the instructions that Chris posted did not solve it, so had to re-start with a fresh DB.
AOB
Next BiLD in 2 weeks.
LHCbDIRAC
- Marko Petric will start within LHCb officially on July 1st
- M2Crypto:
- 1 more machine, maybe one issue found
- won’t create a PR until it’s certified, will keep hot-fixes for now
- Client using M2Crypto: we should try out ourselves first, in production
- Philippe tried it out
- 1 more machine, maybe one issue found
- BKK
- Cert instance (INT12r) moved to 19c
- need to update the instant client on the machine (in puppet) [Chris]
- The prod server hardware might be upgraded already next week
- The prod server will be upgraded to 19c on July 1st
- Federico will run tests (including performance tests)
- Cert instance (INT12r) moved to 19c
There are minutes attached to this event.
Show them.