MACALLSTR%vax1.physics.oxford.ac.uk@CS.UCL.AC.UK (12/20/86)
What I do, when running long MONITOR jobs, is to run for a specific time(1 day) append output file to the total file and then restart for another day's interval. As the APPEND/COPY/RENAME can be going on in parallel with MONITOR, there is at most a few seconds' interruption in the statistics gathering. Re MONITOR in general. I've seen some very strange figures coming back from MONITOR on our 780 : percentages often greater than 100 ! This occurs both with long-term stats gathering and running on-line. Some other figures appearing are physically impossible.