Choose timezone
Your profile timezone:
PASTE meeting
→
Europe/Zurich
2-R-014 (CERN)
2-R-014
CERN
Description
CERN, bat 2 R-014
To join the telephone conference, click here: https://audioconf.cern.ch/call/0165202
To dial in to the conference:
a. Dial +41 22 7676000
b. Enter access code 0165202
PASTE Meeting, 15th February 2011, CERN
Chair: A.Tsaregorodtsev
Minutes: F.Stagni
CERN: F.Stagni, R.Santinelli, J.Closier, E.Lanciotti, K.Ciba, Z.Mathe, P.Charpentier, P.Clarke, R.Santana, M.Rashid, A.Zhelezov, M.Ubeda
Phone: A.Tsaregorodtsev, R.Graciani
We didn't go through Savannah tasks today. We discussed the following points:
# Production issues
- MC production
NTR
# DIRAC Release Update
- v5r11 patches
There was a patch yesterday, requested by Federico, for the possibility to handle XGEN file type (new MC file type).
Yet another patch for LHCbDIRAC and DIRAC is requested to fix the issue with requests. This is a necessary patch since otherwise the productions will not complete. The tag will be done today by Ricardo. Such fixes should also go in the next release.
- v5r12 release status and certification tests
Joel: fixes requests should go through savannah. This is important so we don't lose track. Andrei: it might slow down the process, but in principle I agree.
Certification is started. Some problems have been shown, but the bulk of the tests still have to start.
There was a bug in the service used to update the software. This has been hotfixed in order to make this tool usable.
Yesterday we had a problem with the web portal, should be fixed now.
- Certification procedure
The proposal by Federico is to launch these 3 "workflows":
- A MC + MC merging (LHCb) workflow
- A reconstruction + reconstruction merging (DaVinci) workflow
- Few replication transformations, specifically tuned to test eventual new plugins that have been deployed.
Certification of the user interface:
Joel: I think thare are 2 parts, one is what Ganga uses, the second is about the new dirac commands.
The Ganga people will be contacted by Joel: they should have a test ready, that could be included in the certification process.
Marco Adinolfi has been contacted to know the list of the commands that are used by the DQ shifters.
The logging mechanism for all the commands have been changed
For what regards the scripts logger, the default level for logging will be DEFAULT. For the users have been intruced:
-d = debugging level. Possibilities are '-d', '-dd', '-ddd',
-h explains everything
New level intruduced is the NOTICE level.
The new portal is not yet in the certification instance. It should be put there very soon. There is not a formal way to test it.
In https://lhcb-cert-dirac.cern.ch/doc/_build/html/ there is a draft regarding certification procedure. Everybody has a look please.
Obviously this is very general for the moment, and will have to be updated.
- Deployment of the new release
There is a proposal for the deployment of the new release. It is reported here since not everybody has seen it:
- We decouple the preparation for the power cut from the deployment of LHCbDirac v6r0
- Concerning the power cut, which will affect LHCb voboxes at CERN between 9h30 and 12h00 on Tuesday 22nd, we propose the following:
. On Friday 18th, we change to one the limit of of jobs that pilots can run (assuming this is possible)
. On Monday 21st morning, we stop the matcher and submission of new pilots
. On Tuesday 22nd at 8h30, all voboxes are shutdown, ahead of the power cut
. When the CERN CC intervention is over and the system is back up, we re-enable pilot job submission, with the old limit - i.e. the system comes back to its current state
Since the vast majority of our payloads complete in less than 24 hours, this procedure should ensure that there are few or no jobs running in the system. The impact on all jobs (user and production) is that jobs submitted after Monday morning will wait until Tuesday afternoon. I propose that this procedure is documented and, if successful, is applied in future whenever the system needs to be stopped for a scheduled intervention on the critical hardware.
On this part, it has been reported that we don't need to change the limit on the number of jobs for pilot.
- Concerning the deployment of LHCbDirac v6r0, we propose the following:
. It is felt premature to stop all production activities before having a successful certification of the new software
. When certification is completed (or when only very minor issues remain to be certified), we stop submission of all new productions
. Existing running productions will be allowed to complete. The deployment date shall be chosen and announced based on an estimate of the time needed to complete existing productions
. All productions running in the system should be marked as completed before the migration
. Production requests that have not yet been started must be migrated to the new system
. In order to ensure a smooth migration, all actions needed for the migration should be documented beforehand, with time estimate, and tried out on the certification system
The time window for the migration remains unchanged. The goal remains to have a certified system available by Monday 21st, what changes is the proposal to keep the production running until we can guarantee a rollout date based on successful certification. Note that the rollout cannot slip indefinitely: the new system has to be up and running well before the start of 2011 data taking. For this reason, already now we will not activate any new production requests expected to run for more than a few days, such as the femtoDST production.
We have decided that at the thursday meeting we'll make the decision on when to stop the production requests.
# AOB
volhcb13 is slc4. It is a development machine, not used at the moment, but a second development machine could be useful, since at the moment there is just one.
(LHCb)DIRAC week: the week of the 21st of March is the collaboration week. Another possibility is the week before. Let's decide on the week of the 21st, where the first full 2/3 days could be available for everybody.
There are minutes attached to this event.
Show them.