vdas_20130307
by
Marie Anne Bizouard
—
last modified
2013-03-07 17:37
VDAS meeting
Agenda
1. Announcements
2. Software management
- CVS --> SVN migration: next steps (cleanup?)
- software alignement organization: alignement_v3.pdf
3. ER3 summary
- muti-core machines
- low latency transfer during ER3
- online processing during ER3
Attendance: Florent, Benoit, Loic, MAB, Franco
Minutes
* CVS --> SVN. One tag (last version) missing in 1 package.
Cleaning? lots of packages. The pool of names is shrinking. Consensus to perform cleaning. Will keep "dead" packages in CVS that will remain in read-only mode. Anyone could at any time ask to switch a "dead" package to the SVN archive.
Procedue:
1/ identify owners.
2/ CVS --> SVN only "alive" packages.
3/ keep CVS archive in read-only mode for all "dead" packages
4/ Prepare schedule + documnetation how to use SVN
5/
Franco: using plain commands. Easy script (4 lines). Need a bit more time to check everything is OK.
Conclusion: switch to SVN just after the Virgo week. Few emails to prepare minds.
Benoit: what about scvs?
Franco: with SVN, scvs is not as useful as with CVS. Not a big deal to have a svn-simple package
MAB: no mistery that I do not like scvs because it prevents multi-user developements. Could we take the opportunity of SVN switch to
Benoit: if plain svn commands do what scvs command did, fine for me.
Conclusion: a little helper with commands (cvs, scvs, svn) translation will be useful.
* Software alignement: few corrections in the document. Franco will start working on VirgoPolicy v2r6 next week. MAB will check it in Lyon. Alberto will be asked to do the same at CNAF. VirgoPolicy/v2r6 should be ready. Root v5r34p5 should be the baseline. Only worry: SL4. Franco will continue tests. If pbs can not be solved easily a solution would be to maintain a special root/v5r26 Dy version for Cascina SL4 control room obsolete machines.
virgoStaging: Emmanuel made modifcations in Pm that broke things. Testing phase (sandbox). Franco will send an email to explain how this works in the coming weeks.
* ER3 summary
Benoit: multi-core machine working fine, but no time to work a lot on MBTA. Need a multi-thread version ....
OK to share the machine for omicron tests.
Low latency transfer: bandwidth limitation when many channels were sent.
Florent : omicron 3 channels. Run smoothly. Now move to run
* AOB: next VDAS call: the week after Maryland meeting
1. Announcements
2. Software management
- CVS --> SVN migration: next steps (cleanup?)
- software alignement organization: alignement_v3.pdf
3. ER3 summary
- muti-core machines
- low latency transfer during ER3
- online processing during ER3
Attendance: Florent, Benoit, Loic, MAB, Franco
Minutes
* CVS --> SVN. One tag (last version) missing in 1 package.
Cleaning? lots of packages. The pool of names is shrinking. Consensus to perform cleaning. Will keep "dead" packages in CVS that will remain in read-only mode. Anyone could at any time ask to switch a "dead" package to the SVN archive.
Procedue:
1/ identify owners.
2/ CVS --> SVN only "alive" packages.
3/ keep CVS archive in read-only mode for all "dead" packages
4/ Prepare schedule + documnetation how to use SVN
5/
Franco: using plain commands. Easy script (4 lines). Need a bit more time to check everything is OK.
Conclusion: switch to SVN just after the Virgo week. Few emails to prepare minds.
Benoit: what about scvs?
Franco: with SVN, scvs is not as useful as with CVS. Not a big deal to have a svn-simple package
MAB: no mistery that I do not like scvs because it prevents multi-user developements. Could we take the opportunity of SVN switch to
Benoit: if plain svn commands do what scvs command did, fine for me.
Conclusion: a little helper with commands (cvs, scvs, svn) translation will be useful.
* Software alignement: few corrections in the document. Franco will start working on VirgoPolicy v2r6 next week. MAB will check it in Lyon. Alberto will be asked to do the same at CNAF. VirgoPolicy/v2r6 should be ready. Root v5r34p5 should be the baseline. Only worry: SL4. Franco will continue tests. If pbs can not be solved easily a solution would be to maintain a special root/v5r26 Dy version for Cascina SL4 control room obsolete machines.
virgoStaging: Emmanuel made modifcations in Pm that broke things. Testing phase (sandbox). Franco will send an email to explain how this works in the coming weeks.
* ER3 summary
Benoit: multi-core machine working fine, but no time to work a lot on MBTA. Need a multi-thread version ....
OK to share the machine for omicron tests.
Low latency transfer: bandwidth limitation when many channels were sent.
Florent : omicron 3 channels. Run smoothly. Now move to run
* AOB: next VDAS call: the week after Maryland meeting