Status of PDC’06 Latchezar Betev TF meeting – September 28, 2006
Status of PDC’06 2 Weekly report (1) The problems reported at the last TF meeting (APIService and SEs overload) have been partially solved: APIService is working fine with 2 load-balanced machines Distribution of load on several SEs helped This is not a final solution (rather a workaround) The xrootd behaviour is now monitored by the AMON service of AliEn and restarted as necessary Since two days we are processing steadily at an average of ~1700 concurrently running jobs
Status of PDC’06 3 Weekly running profile SEs stable
Status of PDC’06 4 Weekly report (2) Typical problems Prague (Dagmar) VO-box proxy expires without warning grid- and myproxy- still valid GridKa – proxy problem on a system level SARA – proxy problem in the VO-box Several sites have other operational problems, which are being solved Houston – NFS problems Kolkatta – cluster shutdown Several sites (Krakow, Poznan, LBL, Mexico) are being installed and debugged Good news GSI and Lyon (problem with package installation) are back in action
Status of PDC’06 5 Weekly report (3) We are now in a normal operation mode The available resources are saturated with running jobs Awaiting the new AliEn release (v.2-12) Details from the AliEn experts New AliRoot release v4-04-Rev-09 is in the final stages of validation New implementation of OCDB access classes Object names retrieved from metadata search – one connection to APIService for all objects (reduction of x64 of access frequency) Jobs which end in ‘segmentation violation’ will no longer hang on the WN doing nothing
Status of PDC’06 6 Data statistics Special productions Di-muon 100K (done) p+p 900GeV 200K (done) Single-muon 100K (processing) Standard p+p min.bias: 1MiO events in the last week, 800K of these in the last 2 days We hope the production will continue in this way