Global Calibrations
Global calibrations are done regularly, from the Run Control GUI at
disposition for the DAQ shifter.
The GUI looks like this
The calibrations are performed on dedicated farm nodes, bbr-farm41-45.
The correspondence is the following:
SVT : bbr-farm41
DCH : bbr-farm42
DRC : bbr-farm43
EMC : bbr-farm44
IFR : bbr-farm45
The logfiles for Run Control for the calibrations are in
/static/logfiles/Orc/$SUBSYS/
and you are interested mainly in files :
SM.log-MMDDHHMMSS for list of actions taken by any process
involved in calibration
dataflow.log-MMDDHHMMSS , specific of dataflow process
How does run contro start calibrations ?
At the moment, run control is launching subsystem provided scripts,
for running eventlevel and for running validation. That's what's
running when you see the subsystem botton on calibration GUI going
into states CALIBRATING and VALIDATING respectively.
Run control starts taking calibration data after a "sleep" period set
in the eventlevel scripts, or after reading sentence "Oep eventlevel
successfully connected to partition" in the subsytem calibration
logfile (see subsytem specific script).
With the present implementation of OEP, this is the only way to do
it.
You can find the scripts run control uses for calib/valid in the Orc
directory of any online release, e.g.
$BFDIST/online_releases/3.2.23/Orc/${SUBSYS}_EVENTLEVEL
$BFDIST/online_releases/3.2.23/Orc/${SUBSYS}_VALIDATE
and from there you can trace what your subsytem is
running(scripts,tcls).
Some minimal requiremnts for the scripts (eventlevel/validation):
1) if some error happens in the execution of the set of instructions
you issue in the script, or if the validation fails (e.g. compared to
reference, data are not good enough), then you should exit the scripts
with a NON-ZERO exit code.
ZERO exit code means for run control that everything went fine , or
validation was ok.
2) do NOT hardwire specific releases executables in your script. Use
the /nfs/bbr-srv02/bfdist/Production/bin/SunOS5/findOnlineProdExec
script by Gregory (gpdf@slac.stanford.edu) to find the right event
level executable to run.
3) Try to put as little connection to AFS as possible, since we are
trying to move to a system AFS-outage proof. Talk to Steffen Luitz
(luitz@slac.stanford.edu) for getting an area on NFS for your specific
scripts, and when done, ask the run control expert (franek@slac.stanford.edu)
to change the run control scripts accordingly.
Please, subscribe to run
control HN and post there problems/questions concerning calibrations. The expert will answer you as promptly as
possible.
xella@slac.stanford.edu
Last modified: Mon Oct 9 17:32:07 PDT 2000
|