Choose timezone
Your profile timezone:
PASTE meeting
→
Europe/Zurich
Description
CERN, bat 2 R-030
To join the telephone conference, click here: https://audioconf.cern.ch/call/0122882
To dial in to the conference:
a. Dial +41 22 7676000
b. Enter access code 0122882
Production and PASTE Meeting, 20th February 2007, CERN
Chair: A.Tsaregorodtsev
Minutes: S. Paterson
CERN: M.Bargiotti, G.Castellani, P.Charpentier, J.Closier, A.Smith, S.Paterson
Phone: A.Carbone, C.Cioffi, N.Brook, R.Graciani, A.Tsaregorodtsev
1. Production Status Report ( J.Closier )
Minutes from the Monday meeting are available here. Joel reported that all jobs for the new productions and the
jobs correcting BUG#1 have been submitted. Productions correcting BUG#2 have been prepared.
Stuart reported the current progress with testing the new job wrapper and lcg-getturls. CNAF, GRIDKA and IN2P3
worked ok but the remaining Tier-1 sites had problems. Ricardo spotted that the permissions were not correctly set
for the DCAP libraries which caused the application failures, this should be resolved in the software distribution tarball.
( ACTION: Joel ) Tests are to continue and if successful the new job wrapper will be put into production. ( ACTION:
Stuart )
2. Production Operations Issues
Andrei reported several fixes are in place for Data Management affecting the Transfer Agent and the Bookkeeping
client.
Ricardo mentioned that transfers to Tier-0 tape should not be part of the transfer request from now on, Andrei
confirmed this is the case for the latest DIRAC release. Another issue was to prevent source and destination SURLs
from being the same, Andrei agreed to look at this. ( ACTION: Andrei ) Ricardo reported that some transfers between
PIC and RAL had been affected by a problem with one of the GridFTP servers, this has been identified and will be
resolved.
Another issue raised by Ricardo was that SE's block requests in case of failures (leaving threads blocked for several
hours). After some discussion it was agreed to limit the number of threads per 'channel' in the logic of the Transfer
Agent. ( ACTION: Andrei )
The gLite Agent Director thread should be put into production however Andrei reported that no pilots were sent for
submitted jobs thus far. The default mode of operation should be for pilots to be sent to both LCG and gLite RBs
which requires monitoring of the gLite pilots through the Agent Monitor service. This should be developed on the test
system before putting the gLite Agent Director into production. ( ACTION: Gianluca )
Philippe mentioned it would be advantageous to be able to monitor different job types (e.g. production, processing)
through the DIRAC monitoring pages. Ricardo commented that major improvements are planned for DIRAC3 but
agreed to put in place a new overview for the pie chart and querying by job type in the historical plots. ( ACTION:
Ricardo )
The issue of Accounting was also discussed and it was suggested that certain plots e.g. execution time / CPU
consumed should be automatically generated. Proposals for other plots are to be circulated to the developers list.
Philippe raised the point that dashboard, Raja's monitoring page and the DIRAC monitoring appeared to be
inconsistent. Andrei reported that Raja's page was based on inaccurate information and should be scrapped (
ACTION: Raja ) and It was later determined that dashboard information is gathered over a small time period which
offers an incomplete picture. The confusion over timings should be cleared up and the complete list of RBs currently
in use should be communicated to the dashboard developers. It was also decided to report the activity type (e.g. test,
production) and application status to dashboard, potentially using Mona Lisa. The ATLAS approach should be
investigated to facilitate this. ( ACTION: Gianluca )
For the data removal and LFC clean up Andrew reported that the definite set of data to be removed has now been
determined. The list of directories will be sent to the relevant site admins for the removal of physical replicas. (
ACTION: Andrew )
Marianne reported that she would email the LFC developers by the end of the day to start the SE name substitution.
Andrei requested that the list of deprecated SE names still to be tidied from the CS would be sent to him once this is
completed. ( ACTION: Marianne )
Stuart reported that protection had been added to the job wrapper to cope with LFC entries containing a whitespace.
The affected entries should be identified and fixed. ( ACTION: Marianne )
3. Bookkeeping on volhcb01 status (C.Cioffi)
Carmine reported that the migration of the bookkeeping to volhcb01 was essentially complete. After the webpage has
been migrated to AMGA and tested (timescale 1 week) it was agreed that lbnts2 can be retired.
4. DIRAC3 Tasks
Joel has prepared the DIRAC3 directory structure, it was agreed to remove Information/ and have e.g.
AccountingSystem on the level above.
Ricardo reported that the basic functionality for DISET, the logger and the CS was in place for the new framework.
The next week's work will be on integration with the new directory structure.
For the build tools Ricardo pointed out that the DIRAC python code would also have to reside in an architecture
specific directory to allow shared installations such as the DIRAC API on lxplus to work.
5. AOB
None
There are minutes attached to this event.
Show them.