Re: Keeping clean pages in shared memory

> We wanted to know this because we didn't know if page cleaning and > checkpoints had a detrimental effect on write caching. It appears that > it doesn't so we felt we could change the LRU min and max dirty to a low > value and not affect performance. Not quite David, if you flush a buffer too often you could harm performance because: 1) If a buffer is written to, say, every 30 seconds and you flush it once a minute, you'll do five times as many disk writes for it than with a 5-minute checkpoint 2) If the disk heads "wanted" to be somewhere else at the time of the fuzzy flush, you'll increase head movement, albeit maybe not very much. 3) A nice big checkpoint every five minutes, with sorted writes, is likely to have a lower head-movement-to-pages-written ratio than a small fuzzy flush every minute. Couple of grey areas though. - Are critical sections suspended during a fuzzy flush? (If they are then my head contention argument is less strong) - I can't recall whether fuzzy flushes do sorted writes. I seem to remember reading one manual or document that said they do, another that said they don't. - Does anyone know how the write sorting mechanism works? What overheads does it incur? Does there come a point when the sorting routine could itself requires temporary disk writes? I've heard tell of writes sorts being quite slow sometimes. What factors determine whether they're a good thing? akent@cix.compulink.co.uk (Andy Kent) ------------------------------------------------ Freelance Informix Database Specialist, Redland, Bristol, England