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
Production and PASTE Meeting, 12th January 2010, CERN
Chair: A.Tsaregorodtsev
Minutes: A.Tsaregorodtsev
CERN: F.Stagni,J.Closier,A.C.Smith,V.Romanovsky,R.Santinelli,Z.Mathe
Phone: A.Tsaregorodtsev,A.Casajus,M.Sapunov,A.Zhelezov,E.Lancioti,A.Zhelezov
0. Ongoing Action List
10919 - 10 machines are delivered and configured in the same way as volhcb12. Those are ready to
host the production DIRAC services.
12526 - LogSE host is available but its configuration is still ongoing.
1. Production issues
6 Productions defined and accomplished in January, nothing special to report.
3. Operational issues ( Ph.Charpentier )
- Tests of the DPM sites.
The tests were successful with the data in the CPPM-USER in Marseille with no special effort.
The libshift.so to libdpm.so link was there after the software was installed. This should be
confirmed also by tests in Edinburgh.
- Job information in the BK. Joel reported that the Bookkeeping module is updated to report properly
the job information. Still to be checked.
- Avoid migration of dirac_directory files to tape. Andrew reported that the use of dirac_directory
files will be completely removed in one of the January releases.
- Decision of keeping core services at PIC or moving to CERN and have PIC as backup. Elisa reported
the network reconfiguration at PIC will be updated to hopefully reduce the latencies especially
for queries coming from CERN. The DNS settings at PIC will be done such that queries will be
redirected to Barselona server in case of the PIC web server inavailability. This should reduce
response times. Once this will be done we will measure the performance and take the final decision.
However, this does not exclude another Web Portal at CERN which will still be maintained for redundancy.
The Accounting and Monitoring services in Barcelona University will continue. All the servers
can be accessed by the DIRAC managers via gsissh in case intervention is needed.
- FTS + WMS client middleware available in LCG-AA. Ongoing.
- Keep real data on disk at CERN (RAW and RDST) as in the Computing Model.
- Check availability and test usage of xrootd at sites. It was discussed that LHCb will not ask
for direct access to xrootd servers bypassing the SRM TURL resolution mechanism. If SRM at a site
is configured to return xrootd URL, it can be used. However, there is no real preference to use
xrootd or other remote file access protocols. In particular, IN2P3 is making access through dcap
protocol ( as opposed to gsidcap ) which should make access more reliable.
- Platform detection and capability. This is being addressed by the small working group consisting
of (Hubert, Joel and Stuart). Adri will also join the discussion.
- Use LHCb release tools for clients on AFS. These tools were ready, prepared by Joel. However, the recent
renaming of LHCbDIRAC project has screwed up the DIRAC release tools. During the meeting Andrei
reported that the last DIRAC release building did not work. Right after the meeting, the problem was
identified as related to the LHCbDIRAC project renaming. So, we still have to define the
exact release procedure which is suitable for both DIRAC services and AFS client installation.
To be discussed and decided by a working group ( Andrei, Adri, Joel, Philippe, Marco ).
- Distribution of RAW data to Tier1s by runs. Ongoing as reported by Andrew.
- Changes in the BK (in synch with Online) for the configuration names and DAQConditions, as defined in December.
Zoltan has reported that the necessary changes are already done and will be available in the coming release.
The corresponding Savannah tasks will be formulated if not yet there (ACTION: Andrei).
4. DIRAC releases
DIRAC v5r0p0-pre2 was compiled. However, the release preparation failed because of the recent
LHCbDIRAC project renaming, the release tools should be restored. Once done the release will be
ready for installation on the Certification Setup.
5. DIRAC tasks
- SVN based release
The new SVN code structure based DIRAC codes are extansively tested in the Test Setup. Few
release have been compiled to support these tests. Now we are preparing the release to be
installed on the Certification Setup. The system packages were tagged, the overall release
building failed, to be sorted out (After the meeting the problem was identified as due to the
LHCbDIRAC project renaming).
- The hardware for the Certification and Production setups is availbale
- The service host installation scripts are being finalized, in particular the more versatile
MySQL installation and database setup scripts. Ricardo is finalizing it. The goal is to have
a one command host installation which eventually can be included into CERN/FIO standard
templates.
- The recent introduction of Secure Linux features on the VO-boxes was creating many
problems that were difficult to understand due to a lack of documentation. In particular,
making ports closed even if accessed through the local loop was preventing the Apache/Pylons
installation on the Development system.
- once the host installation scripts will be finalized, we will go through Certification
chain of prereleases, before production release.
Subsystem status and news
- Monitoring (SAM, SLS, Dashboard). Roberto reported about the update of file access tests within SAM framework,
and incorporation of the Andrew's unit test for the SRM interface. The tests are now being migrated
to the new DIRAC releases framework.
- DIRAC Accounting and Monitoring. Adri reported that the new SVN based Accounting and Monitoring
services in Barcelona are successfully tested. The Accounting is not using the new Plotting package
of DIRAC. This was considered to be not very urgent.
- ReseourceStatus service. Federico reported that he has added SE monitoring to the sensors, transfer
quality monitoring, in particular. LFC monitoring will be added soon. Federico is now using
Notification system to send notifications and alarms. The use of alarms should be better understood.
In particular, the alarm criticality level should trigger different actions. Federico and Adri
will look at how this can be implemented and used.
- WMS. Ricardo has provided a command to extract job related information provided gLite pilot reference
and the time stamp. This is needed for the user job traceability in the long lasting MUPJ saga.
- Web Portal. Adri has reported that the separation of the DIRAC generic Web portal stuff and LHCb
specific pages are done and supported by the installation tools. Andrew suggested to have a possibility
for system specific Web portal pages, which Adri will consider.
- Documentation. The Sphinx/RST based documentation pages produced by Federico are still to be evaluated
by other developers. Adri offered to write the Web Portal installation instructions using these
techniques.
There are minutes attached to this event.
Show them.