Re: Mirroring

Jay Aymond wrote: > > "Art S. Kagel" <kagel@bloomberg.com> wrote: > > >Jay Aymond wrote: > >to wait for page flushes during checkpoints. I run my busiest servers > >with > >LRU_MAX_DIRTY=3 and LRU_MIN_DIRTY=0 90% LRU Writes, these puppies fly! > > > >Art S. Kagel > > Thanks for the info ---- I will give it a try. > > Is there some rule for defining how many LRU queues to have for a > certain number of buffers? No. Unfortunately, it is trial an error and tuning and watching your stats. Look at how long your checkpoints are taking (hard in 5.0x you have to guess from the checkpoint times which won't work if PHYSLOG overflow is causing unscheduled checkpoints). Long checkpoints (you define long for us its >7 seconds for many <1min)? Shorten LRU_MAX_DIRTYand LRU_MIN_DIRTY and/or increase the number of LRUS and CLEANERS so that fewer buffers are dirty, on average, at checkpoint time. Too much Chunk Write I/O? Lengthen checkpoint interval and increase PHYSICAL LOG size to reduce the number of checkpoints. Fg Writes? More buffers, more LRUs and cleaners. You get the idea. Good luck. Art S. Kagel