Production and PASTE Meeting, 15th September 2009, CERN
Chair: S.Paterson
Minutes: S.Paterson
CERN: P.Charpentier, J.Closier, Z.Mathe, S.Paterson, M.Rashid, R.Santinelli, A.C.Smith, V.Romanovsky, F.Stagni
Phone: A.Casajus, R.Nandakumar, M.Sapunov, A.Zhelezov
Apologies: R.Graciani, A.Tsaregorodtsev
0. Ongoing Action List
SLC5 testing, new binaries should be deployed today or tomorrow as many sites are migrating already.
- Production request page update
Alexey is waiting for a response from Matvey about how to open certain productions in the monitoring page. A more
optimized BK query in order to obtain the number of events would be advantageous.
1. Production Status Report
- Report from the Production meeting
The recent SSL Handshaking errors affecting the Matcher caused a ramp-down of running jobs and should be
investigated. ( ACTION: Adri ) Apart from the 10% failure rate still happening with Gauss, production is running
smoothly.
- Set up stripping workflow(s)
Ongoing pending DaVinci release.
- ELOG deployment
The suggestion to deploy the ELOG on one of the offline machines was discussed, during the meeting Olivier sent a
mail conveying this was understood. It was decided to continue with the current situation given that we now have a
support mailing list to rectify problems.
- Production status updates
With the developments from Stuart and Andrew, there are two new agents governing the production status machine,
documentation for this will be prepared. ( ACTION: Stuart )
2. DIRAC Release Update
The status of the v4r19 candidate release is not clearly understood. There was some discussion regarding the SVN
migration and we were pending updates from Marco regarding the structure.
3. DIRAC Tasks
- Infrastructure monitoring (SAM,SLS,Dashboard)
Roberto mentioned that one test has been demonstrated in the Nagios infrastructure. In addition there is an alarm
system for available storage space at the Tier-1 sites (excluding T1D0).
- Bookkeeping status
Zoltan implemented a method to summarize the events for a given event type and also the productions for a given
configuration name and version. Philippe wondered where we are with the BK documentation and Zoltan will add the
link to the 'help' button. Matvey mentioned the progress with the BK web interface, some cases should be tested with
Zoltan.
- DMS status
Recent work has been on the data integrity tools. The integrity client can now perform all possible checks and
methods to resolve different kinds of problematic files are being provided. This is ongoing as a decision tree may be
implemented to streamline the code.
- Production Management status
Production status machine has been modified as discussed above.
- VO-boxes
Raja mentioned that things are stable on the VO-boxes.
- Web Portal
A certification instance of the web portal would certainly be useful in order to improve the release cycle. For now a
machine in Barcelona will be used.
- Services deployment ( hardware, configuration )
Raja mentioned that the VO-boxes will be updated to SLC5 with a medium priority so we should be ready for this.
4. AOB
The issue of testing pilot submission through the CREAM CE was raised by Philippe. After some discussion it was
agreed that submission to the CREAM CE should be demonstrated at least as a test. ( ACTION: Stuart, Ricardo )
Several LHCb people have been added to the WLCG Technical forum list. The intention behind this group is to identify
major problems from experiments and sites and convey these to the middleware developers.
Preparations for the EGEE demonstration in Barcelona are ongoing, Roberto has a movie from Stuart for production
requests. For the Ganga part there will be some coordination with Andrew Maier.
Philippe reminded that there will be a mini-workshop on virtualization scheduled for the Friday of the software week,
internal to LHCb.
There are minutes attached to this event.
Show them.