Re: RA, bugs, and performance

Posted in 1998

Art S. Kagel — Mon, 15 Jun 1998 15:08:59 -0400 — source: Informix-list mailing list archive (1991-1998)

Arun Shastry wrote: > > Andreas Zeugswetter wrote: [SNIP] > > Hmm, what you say can only be true on a system very low on memory. > > Our values are RA_PAGES 256 and RA_THRESHOLD 96 and we get very [SNIP] > See if you get any different performance, both query and cache, if you set them > to > lower values. I'm curious. I'd put money down that his read cache % is lower than it will be when he reduces his RA_ values! Unless he is using REALLY old legacy AT class (80ms 2400 RPM) disk drives or a large farm of singleton drives! Naw couldn't be not for a DSS system! Art S. Kagel