we have observed that when someone sends a job that produces very large log files in the workers, because he forgot to remove debugging printouts, later during the merging process when the master collects all the logs from the different workers, this merging process crashes and the proof daemons are destroyed. I have correlated this crashes with large sizes of log files produced.
Does anyone know a way to protect a proof farm in this kind of situations?
Any configuration parameter that establishes a limit of the size of the log files being collected?
thanks for the help, Carlos