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, 25th January 2011, CERN
Chair: A.Tsaregorodtsev
Minutes: F.Stagni
CERN: F.Stagni, R.Santinelli, J.Closier, E.Lanciotti, V.Romanovsky, K.Ciba, S.Reuser, Z.Mathe, P.Charpentier, P.Clarke
Phone: A.Tsaregorodtsev, A.Zhelezov, R.Santana, A.Casajus
Apart from some Elisa's tasks, we didn't go through Savannah tasks today. We discussed the following points:
# Production issues
- MC production replication issue
Vladimir is waiting for instructions on what to do. Some files were marked as "unused" and they got replicated.
We discussed the possibility to select which plugin to choose. This will require also modifications in the production API, and in the templates also, and Federico will take care of them. These modifications should go in the patch release that we're preparing.
Philippe agreed on writing a small tool that checks the replication of the files. A savannah task has been opened.
# Data Management issues
- Storage Usage Accounting
Adria fixed the storage usage DB. The Storage usage agent now removes empty directories also by the catalog. It is installed in a test machine in Barcelona (Adria sent a mail around few days ago), but he'll now deploy it the standard LHCb test machine. The functionalities provided are available to go into the next pre-release, to be certified.
- Data archival
The last week Philippe discussed about an operational procedure, that will not need much development. Anyway, if needed we have first to decide exactly what we want to do (there are good suggestions by Ricardo) and who will make the develoment: probably Elisa. The replication plugin will need to be modified.
The whole proposal is the twiki https://twiki.cern.ch/twiki/bin/view/LHCb/ArchivingDatasets
The procedure still needs a bit of elaboration. The technical way to achieve it (e.g. having just one replication production or 2) could be discussed in the future.
# DIRAC Release Update 20'( )
- v5r11 patches
The patch release has been delayed for quite a while and now it's time to deplot. Federico will do the modifications to enable the possibility to choose between different plugins (by modifying the production APIs and the templates). Philippe is worried by the modifications (the hot-fixes) that have been put into the production system and that might not be in the next patch release. This afternoon the new patch release will be compiled.
We have decided to go through the certification system because the modifications going in this patch release are quite a lot. Also, FTS transfers will be tested.
- v5r12 release status and certification tests
There are discussions about how many releases to make. Andrei said that in principle there are new features that could go now in a release without waiting for the right "window". These possible "intermediate" release could go now in certification. Joel and Peter would intead wait for the next "big" release. After long discussions, it has been decided that a release of DIRAC could be made before the LHCbDIRAC v6r0 release, but it will not be deployed in production before 3 weeks.
NewBookkeeping: Zoltan is (almost) ready, there are just few technical issues that still need to be solved.
WMS: advancements in the submission to CREAM CEs. Accounting, monitoring, everything should be sorted out. Tested on the develoment system, should be tested also in the certification.
- Certification procedure
DIRAC certification testings is a field that needs developing. The certification test that we have now creates 3 production. What is missing is an automatic checker, because right now we make a lot of checks "by hand". We don't think that there will a lot a work already automated in 3 weeks time, but what we can have is a recipe to follow.
An example of such tests regards a general test of the production system. If Federico will write these tests or not, working closely with a "certification manager", is a matter of details: what Federico, as developer will have first of all to do is write down a strategy, that could be automated in the future. The certification tests should start from the request, and test all the process involved, the complete "production workflow".
Joel is writing a document that decribes which operations should be taken to certify a release. After this procedure is clear, it should be applied, first manually.
All the developers will sooner or later will be asked to develop tests at different levels, depending on what the certification manager will ask, or by their own ideas.
It has also been discussed the possibility to extend the tests to the cases of the user level functionalities: command lines and APIs. Joel has already made some work in this field, and will continue on this path.
Together, certification of the production and user activities, will form a (almost) complete suite of tests.
# AOB
DIRAC week: temptatively we can try to have it the week of the LHCb software week. Andrei is going to make a doodle.
There are minutes attached to this event.
Show them.