BiLD-Dev
Bi-Weekly "Loyal" DIRAC developers meeting. And, following, the LHCbDIRAC developers meeting.
Join Zoom Meeting
https://cern.zoom.us/j/91083694183?pwd=ZkdDY1I1YkJVc2o3UTdBY1BRZE15UT09
Meeting ID: 910 8369 4183
Passcode: 12345678
One tap mobile
+41432107108,,91083694183# Switzerland
+41315280988,,91083694183# Switzerland
Dial by your location
+41 43 210 71 08 Switzerland
+41 31 528 09 88 Switzerland
+41 43 210 70 42 Switzerland
+33 1 8699 5831 France
+33 1 7037 2246 France
+33 1 7037 9729 France
Meeting ID: 910 8369 4183
Find your local number: https://cern.zoom.us/u/artUfAdNB
Join by SIP
91083694183@188.184.89.188
91083694183@188.185.118.153
Join by H.323
188.184.89.188
188.185.118.153
Meeting ID: 910 8369 4183
Passcode: 12345678
BiLD (Bi-weekly DIRAC Development meeting) – 08/07/2021
At CERN: Still, nobody :/
On Zoom: Federico, Abderahaman (LUPM new collaborator), Alexandre, Andrii, Christophe, Christopher, Daniela, Hideki, Luisa, Janusz, Simon, Ueda, Xiaomei
Apologies: Andrei
Follow-up from previous meetings
- Ran mini hackathon last week on py3 DIRAC:
- trello board here
- The hackathon was held on the “oauth” certification machine (see https://github.com/DIRACGrid/DIRAC/wiki/Certifications for details)
- py3 server installation
- dteam VO via VOMS
- no ES, no VMDIRAC
- Mostly checked WMS + DMS bits (e.g. TS not tested)
- PRs merging strategy?
- this is done, but some discussions emerged so not yet applied (see PR). It would be about time to conclude
- “parallel” discussion on using
blackfor formatting- Q: is it really necessary?
- “parallel” discussion on using
- this is done, but some discussions emerged so not yet applied (see PR). It would be about time to conclude
DIRAC communities roundtable
GridPP:
Daniela
- Running v7r1p39
- v7r2 on the pre-production server
LHCb:
Federico+Christophe+Christopher
- NTR
Belle2
Hideki+Ueda
- DBs extension: TaskQueueDB, TransformationDB
- TransformationDB and Rucio integration likely problematic? Will open an issue on GitHub
IHEP
Xiaomei
- v7r1p30 in pre-production, going to move to it in summer maintenance period
- WebApp extension: need to use the docker container for creating releases
- Compile just the WebApp (for development): https://dirac.readthedocs.io/en/latest/DeveloperGuide/WebAppDIRAC/DevelopWebApp/index.html#create-your-first-example
- Making WebApp releases: https://dirac.readthedocs.io/en/latest/DeveloperGuide/ReleaseProcedure/index.html#deploying-dirac-tarballs
CTA
Luisa
- Instance distributed between IN2P3 and PIC
- MySQL transfer (MariaDB 10.5)
GROUP BYstatements change- PR will need to be make, will be backward compatible
- Running v7r2p11
- tested “almost” everything
- Asked by the ESCAPE project to participate in the next DC, which is using Rucio
- asking help from Belle colleagues, in August
- Rucio version in DIRACOS==1.25.5
DIRAC releases
- v7r1p44:
- NEW: (#5201) Implemented a function that allows programmatic listing of directories in the DFC.
- CHANGE: (#5211) FileCatalogClientCLI register File now takes a fixed number of arguments (6) including Checksum
- addresses RegisterFile in dirac-dms-filecatalog-cli have no checksum input #5203 which is not closed…?
- v7r2p12 + v7r2p13:
- NEW: (#5212) SiteDirector: added flag for pilots to install python3 DIRAC clients
- CHANGE: (#5219) Service Handlers discover DBs in Extensions
- which is not a full solution (see discussion in PR)
-
- Several other fixes, some for py3
- v7r3-pre13 (7.3.0a13) + v7r3-pre14 (7.3.0a14):
- Several py3 fixes (installation + actual code)
DIRAC projects
DIRAC:
Issues by milestone:
- v7r1:
- Set XrdSecGSIDELEGPROXY=1 to enable third party copy with xrootd solved, closed
- Still there are some old documentation requests.
- v7r2:
- v7r2/r3 multi-core submission does not start multi-core jobs on HTCondorCEs probably “done” (see PR)
- Q: would it require a more recent version of the HTCondor client that we use in DIRACOS? Or is it only for server side?
- Allow for RunningLimit per CE probably also “done” (see (PR, not reviewed)
- DIRAC_USE_TORNADO_IOLOOP set after Refresher is imported in tornado_start_all solved, closed
- v7r2/r3 multi-core submission does not start multi-core jobs on HTCondorCEs probably “done” (see PR)
- v7r3:
- 2 new issues added:
- do not use BLOB in JobParameters MySQL DB table (in general we should remove all BLOBs…)
- [RSS] VOFlag in rss scripts discussion
- 2 new issues added:
PRs discussed:
- [v7r2] Adding GPUs support (CEs definition)
- need review
WebApp:
- Several PRs open+reviewed+merged for py3 (release 4.3)
- 4.3.0a9 on pypi for py3: https://pypi.org/project/WebAppDIRAC/
- To reach there we:
- Created
dirac-webapp-packaging(on github and on pypi)so that WebApps can be automatically built during packaging for PyPI or installation with pip - Created
DIRACWebAppResources(on github and on pypi) to distribute the ExtJS SDK to reduce the size ofWebAppDIRACreleases - The static directories served by nginx are now symlinked to
/WebApp/StaticResourceLinkDirwhile launching the WebApp so the path is predictable - Reflected these changes in the puppet profile (unmerged still)
- Created
- Andrii wants to use the same
tornadoservice for DIRAC and WebApp- as of now being tested on authcertif
Pilot:
DIRACOS:
xroot5?- NTR
condor 9?- NTR
DIRACOS2:
- from previous meeting
fts-reston py3?- NTR
- from previous meeting Andrii jwt + other token stuff need to be added here
- Not yet done
- During the hackthon found that an issue on how
arcwas built, new release soon- would be useful to also test HTCondor (should be OK, given we ran jobs in oauthcertif setup)
VMDIRAC:
- from previous meeting Agreed to merge into DIRAC for v7r3
- PR created: it has been reviewed, and moved back to draft
Documentation:
- NTR
OAuth2:
- from previous meeting Any news from issue WLCG and Token transition – reminders and requirements ?
- There will be meeting again on CE submission
- from previous meeting Andrii forwarded https://docs.google.com/document/d/1ZQzElD866yV6t9AomeW6r-L9s-64pNB95OFDd1EZEbA/edit?ts=60814a09#heading=h.ur0csvxkk8tc which contains technical drawings about the proposed implementation.
- NTR
- Added to oauthcertif the latest code. Andrii will test it himself, getting ready for next hackathons
tornado/HTTPs
- Fix for
tornado_m2crypto(https://github.com/DIRACGrid/tornado_m2crypto/pull/5) merged- Necessary for running tornado HTTPS services on py3: (PR)
- Added 3 WMS HTTPs services in PR for v7r3
management
- from previous meeting 1 task left for deploying on CVMFS
- Not yet done, Marko left, Andrei should take over
- 3 PRs (#40, #41, #42) created and merged for py3
- Also for creating WebApp releases
diraccfg
- Release 0.2.2 created, added a sanity check of the CFG
- and spotted bugs (in 2 tests) following that – PRs for v7r2 created+merged+propagated
COMDIRAC
- NTR
DB12
- (forgot to mention in the meeting… )
- Summer student (LHCb) developing the DB12 package:
- py2/py3
- adding tests, github actions
- publishing on pypi
- using it in Vanilla DIRAC (to do)
- running tests for different MP + other architectures (later)
other externals, including Rucio
- Issue: [URGENT] Rucio MUST be an optional dependency (#5250)
- LHCb developers keep fixing
ruciointegration issues and that’s not cool - Federico it would be enough to have proper responsibility by someone of those that are interested
- Christophe an optional dependency would make sure we don’t need to think about it
- Ueda we have it in BelleDIRAC extension anyway
- not found any reason for NOT making a RucioDIRAC extension
- Cedric’s not present to the meeting making it a bit difficult to take final decision
- LHCb developers keep fixing
- Janusz working for making rucio-DIRAC multi-VO implementation, but most importantly seeing if there’s really the need for a rucio configuration file (it should not).
Release planning, tests and certification
-
Certification machines
- lbcertifdirac70 machine:
- Did not touch it lately
- lbcertifdiracoath machine:
- installed py3 server
- Should move from VOMS to IaM
- lbcertifdirac70 machine:
-
Next hackathon(s)
- On July 15th (who’s not there?)
- Federico We can go back to lbcertifdirac70 and run on py2, do a “standard” hackathon
- For the hackathon in 3 weeks we can move lbcertifdirac70 to py3
- flick back and forth py2/py3? possible but not advisable
- We might also call a mini-hackathon on oauth/token
- On July 15th (who’s not there?)
AOB and topics from Google forum
Next hackathon (py2) on July 15th: https://indico.cern.ch/event/1056182/
Next BiLD on July 22nd: https://indico.cern.ch/event/1057796/
LHCbDIRAC
- https://trello.com/b/BQvEbteg/deploy-v10r2
- No changes in the last 2 weeks
- We first need to install a new release, but for that we’ll first need a DIRAC release
- Won’t install LHCbDIRAC release before Wednesday
- We need new boxes + duplicate the JobManager
- python3 client on CVMFS
- Ganga and PID-calib using it
- Christophe keep bumping in division errors
- after release we can make it the default, also for the pilot
- SandboxStore: installed in prod
- 46% migrated -
rsync-ing - disabled the cleaning
- 46% migrated -
- we should start LHCbDIRAC hackathon based on v7r3/py2 and later py3
- We’ll need to move server by server
- the BKK might need some updates (it’s very delicate)