General Farm Info
The PR work is distributed between SLAC and Padova. SLAC performs the PC
pass and tests, while Padova is responsible for the ER pass and
PostProcessing as well as Skimming for new data only (valid on Apr.20,
2005).
PR Users
- At SLAC the main PR production is done from babaropr account.
bbrobjy is the one which takes care of DB and snapshot
operations. bbrxtc is the one for import/export of xtc and
root files to/from Padova. PR work can be done from any SLAC machine,
although some special actions require log in one fo the farm machine.
The import/export is operated on bbr-xfer02.slac.stanford.edu
- In Padova, PR work is done by bbrprmgr, while xtc operation
are done by bbrxtc. The main machine for PR is
bbr-user02.pd.infn.it and for xtc/DB it is
bbr-import.pd.infn.it.
Farms
- There are 2 PC production and 1 test farms at SLAC:
- PC1: oprserv01
- PC2: oprserv02
- PCTest: oprtest01 (for tests)
Each PC farm has approximately 16 client nodes with double CPU.
They are called tori00XX (valid on Sep.20, 2004).
The names of machines could be found in
/nfs/oprserv0x/u1/PCx/OprConfNodeListPCx.xml
- There are 7 ER production and 1 test farms at Padova.
During Run5, 2-4 ER farms should be sufficient to keep up with
IR2 input. The remaining N-1 ER farms are used mostly for
reprocessing, or in time of severe delays. The last farm does the
skimming for Run5 data.
Each ER is connected with a PP farm, which does merging of root files
and prepare QA root files and the giant postscript file.
- ER1: opr1
- ER2: opr2
- ER3: opr3
- ER4: opr4
- ER5: opr5
- ER6: opr6
- ER7: opr7
- ER0: opr0
(for tests)
Each ER farm has approximately 30-50 client nodes with double CPU.
They are called bbr-farmXXX (or 10.100.2.XX where last XX correspond
to the number in bbr-farm name ) (valid on Apr.20, 2005).
The names of machines could be found in
/nfs/oprx/Reprocessing/XXx/OprConfNodeListXXx.xml .
Directory Structure
The directories for all the farms have the same structure:
- At SLAC:
- /nfs/oprserv0x/u1/XX#/prod
- /nfs/oprserv0x/u1/XX#/logfiles
- /nfs/oprserv0x/u1/XX#/xtcfiles
- /nfs/oprserv0x/u1/TAO
where XX = PC or ER.
- In Padova it is:
- /nfs/oprx/Reprocessing/XX#/prod
- /nfs/oprx/Reprocessing/XX#/logfiles
- /xtc on the OPR server
- $TAO_NS_ROOT
Packages
The following packages are currently needed to run the control
system:
- OprProcessingSys
- OprControlSys
- OprQaUtils
Last modified: Thu Apr 21 11:31:36 CEST 2005