DPM dev enlarged meeting

Friday 31 Jan 2014, 10:00 → 12:00 Europe/Zurich

28/R-015 (CERN)

 

 

- puppet template releases and upgrades

 some feedback from Frederic Schaer, was sending mail to dpm user forum

 wants to avoid double modifications, how to propagate fixes to the puppet templates?

 Andrea: we're starting to use github for this, with the cern-sdc-id account
(Fab: who may have the details of the cern account? steve traylen? )
 

- who is using Puppet? August is coming quickly!
... no clear answer about this. Something is moving.

 

- status of the testbeds... what to do to setup one

 - install a machine, open the dpm ports

 - configure it with epel-testing

 - install dpm and configure it

 - let other devs use it for testing, federating, etc

 - promptly let people know if it breaks

 

 French testbeds (Frederique) have an instance ready at lapp,
not aware of the situation in the grid, suspect it's not effective for the lapp instance

Michel idea is to reuse the test dpm at lal. next month should be in a better shape
(Fab: good! We'll like to check all the testbeds in the HC instance)
 

- Czech testbed? no testbed yet, Valtri can try an installation on a vm. will test ipv6. oliver: also sam has an ipv6 testbed.

 

- Fabrizio: fix the link in the main lcgdm, pointing to koji confuses people
 

- Michel: jerky info from google, containing just one link. Fab: we may want to renew the entire wiki when passing to drupal

 

- VFS component? how's it going? Frantishek will migrate the repo to svn. there are some things with the interfaces of dmlite, will need to separate the interfaces? Fab: we need to interact separately
 

- nfs component: not much to say

 

- drop the GAS component? Has been never used, and now it's broken.

People agree to use metalinks whenever possible instead of GAS

 

- file placement... anything missing?

it's in relcand!

 

- instabilities in dpm adapter... are they over ? Seems so in the trunk testbed

- Deadlock in dpm adapter with xrootd and http... is it over? Seems so in the trunk testbed

 

- Going towards the next release of multiple components.

 

- Next meetings

 https://indico.cern.ch/categoryDisplay.py?categId=4897

 

- Alessandra: issues with the n2n atlas plugin. Fab: who wrote it? noone knows! Fab: seems a config problem, however its implementation was suboptimal for performance

 

- Frederic: release notes for puppet? where? andrea: we will put them into github. oliver: what about a changelog. andrea: in the mailing list too!