At CERN: Federico, Christophe, Zoltan, Philippe, Joel, Milosz, Wojtek, Andre', Marko
On Vidyo: Ueda, Marcelo, Andrei, Imperial College, Luisa
Apologies: Xiaomei
Follow-up from previous meeting
- FTS3 policy for v6r20 should be re-shuffled
- also asked to add a flag for which version to use
- python 2.7.13 can stay (no need for 2.7.14 yet)
- DUW: confirmed in CC-IN2P3. Suggested to use Friday to talk about operational matters with CC admins
Current situation
DIRAC:
- v6r19p4 is the last production version
- Just fixes
- dropped check of pyGSI version
WebApp:
- NTR.
- Andrei would like to discuss the changes Zoltan proposed on deployment - out of this meeting
Pilot:
Externals:
LCG bundles:
- Version v13r0 looks OK, and DIRAC v6r19 depends from that
- This version includes python packages like pyparsing and futures, which should be anyway in the externals. To verify
- still not moved its "development" to GitHub. This would also answer the recurrent question of the content of each release
REST:
- NTR, but Wojciech is using it and looks OK
- quite some similarities with WebAppDIRAC, some code can be ported there
BOINC:
PyGSI:
Documentation
DB12:
- We should make a pip (wheel) package and release it, then import it in DIRAC
Release planning, tests and certification
Release planning:
- v6r17 series
- v6r19 series
- v6r19p5 should be out ASAP
- v6r20 series
- pre5 is out
- PilotsLogging "full chain" verified, modulo the "authentication" (see daily dev focus slides)
- new FTS3 looks OK, but requested Chris to add a flag
Certification process:
- v6r20 certification is ongoing:
- Using trello and slack as usual
- not much progress this week
- The lhcb-jenkins server will be retired in few days
- jenkins-dirac server is replacing it
Weekly development(s) focus
- see slides (Pilots Logging)
- decided to stick with RESTDIRAC, which will need to be updated
DIRAC: current PRs and tasks being worked on
- pyGSI -> M2Crypto: maybe for v6r21
- Made several tests
- We decided that we can remove the legacy proxies part
- Add a flag to use it or not, maybe based on an environment variable
- still lacking the sockets part
- decided to look into what's missing before declaring it not a WIP anymore
- formatting with autopep8 PR: looks OK
- dirac-wms vs dirac-admin scripts: time to change the names?
- Philippe will make a plan for changing names and positions altogether
DIRAC communities roundtable
- ILC: Moved to v6r19 in production
- LHCb: Moving to v6r19 next week
AOB
- Next BiLD-Dev meeting will be held in 3 weeks (November 30)
LHCbDIRAC
- "prod" simlink of LHCbGrid v11r4 done
- Need to update the release doc too
- LHCbGrid v13r0 done
- new jenkins-dirac is used, will be completed ASAP
- Need to make the slaves not visible from outside
- and puppetize the cernVMs
There are minutes attached to this event.
Show them.