Re: Measurements: (was: Raw vs. Cooked files)

Martin Fuerderer schrieb: > Hi, > > >>Why is NOAGE set to 0? Surely it should be set to 1? > > > For this work the machine was dedicated solely to IDS > server. No other applcations were running on the machine. > Therefore NOAGE 1 vs. 0 would not have made much of a > difference as there were no other processes to compete > with IDS for CPU-time. > hmm On Solaris 8 we had to learn the hard way - and we were bitten heavily - that there is a certain competition if - you go with the rule of thumbs # of CPUvps = # of phys CPUs - 1 - and you habe more than 8 CPUs - and you have a lot of cleaners - and you have more than 1 or 2 NETvps - and you do not switch off aging The reason ist that Solaris gives so much preferrence to cleaners in a checkpoint, that this slows down checkpoint in an unexpected way. Just let an run for some time and draw a graph of the accumulated CPU seconds of all INFORMIX processes's to get a picture. Solaris calculated position in the run ready queue like base prio + K * (1/accumulated CPU time) As soon as we set NOAGE 1 the problem that cleaners did not get assigned to physical CPUs was gone. Of course, one can always tune in such a way, that LRU writes do the main part of cleaning ...... :) dic_k -- Richard Kofler SOLID STATE EDV Dienstleistungen GmbH Vienna/Austria/Europe