Re: Q: Mirroring - what is the better way

>>>>> "Art" == Art S Kagel <kagel@bloomberg.com> writes: Art> Art> On this last actually they often do (rumor has it that the HP IO system Art> is much better thanks to Informix engineers). I certainly doubt this... or would be very happy... because then I know that our box is even more superior (is the grammatically correct? :-) than an HP. Art> Rumor aside, Hardware mirroring tends to be fastest on raw read/write Art> tests and most hardware mirror systems include the same optimizations Art> that Informix mirroring does. Since the mirroring done by the O/S is done by the device driver, it doesn't care much whether it's on a cooked or raw device. The cooked (or non-cooked) portion happens a layer above it anyway... Art> OS/software mirroring tends to be Art> slightly slower than Informix mirroring on all of the benchmarks that I Art> have seen but not enough to affect a decision. I'd like to know how'd you set it up... I have yet to see what you claim. Please describe your benchmark in detail... down to disks per controller and any striping that was done. Thx! -- Pablo Sanchez | Ph # (650) 933.3812 Fax # (650) 933.2821 pablo@sgi.com | Pg # (800) 930.5635 -or- pablo_p@pager.sgi.com ------------------------------------------------------------------------------- I am accountable for my actions. http://reality.sgi.com/pablo [ /Sybase_FAQ ]