[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] Debugging Large Log Files

All -

My research group is looking for users with big Condor user log files.

We are studying the use of data mining to troubleshoot large systems.
The basic idea is to read in a log file from a large run, perform
some automatic classification, and produce a diagnosis like this:

    "Jobs where Owner==X tend to fail on machines
     where OpSys==Linux and MachineOwner==Y"

(Truth in advertising: the output is not quite that readable yet.)

If this sounds interesting, and you have some log files that you might be willing to share -- in exchange, you get some results back -- please contact me directly, and we'll talk some more.

Best regards,
Douglas Thain