The logfiles for each farm at SLAC are archived by scripts run by Owen Saxton (saxton@slac.stanford.edu).


Archive Parameters - IN PROGRESS

• The logfiles are gzipped by the control system at the end of each run.
• The archive scripts are run by a cron job, once a day on each farm.
• Any core files are deleted and anything unzipped is gzipped.
• Logdirs are tarred together until the tar file is about (?) big ...

To retrieve the logdir for a run you only need to know what farm processed the run. For now there is a script written by Teela which wraps a script written by Owen. It is run by:

• Create a file (run_file) with the runs you want to retrieve, one on each line.


• ssh as babaropr and cd to the directory where you want the files to be written to. It is important to do this as babaropr so that others can later remove the files.


• ~teela/scripts/retrieve_logfiles.sh < farm_name > < run_file > < option >
  where farm_name = e.g. ER5, and option = -a (retrieve all logdirs for each run) or -l (retrieve only the latest logdir for each run)


• This will retrieve tarred blocks of logdirs from HPSS and then extract the requested logdirs from them. The tarred blocks will stay on disk (so they don't have to be retrieved again for later requests) but when you are sure you are done retrieving files in one range you can delete them.


• You can also run Owen's script by hand if you just want to retrieve one logfile or if you want to explore the options.
  ~saxton/OprLogArchive/rlrecov
  call it without arguments to get help.