Re: Data Buffer Cache

"Nuno Souto" <nsouto_at_optushome.com.au> wrote in message news:dd5cc559.0209152049.43a01401_at_posting.google.com...
> "Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message
news:<3d84e994_at_dnews.tpgi.com.au>...
>
>
> > accuracy of the information that was being quoted that's stuffed. If
this
> > was really the way that things happened, the database would grind to a
halt
> > at every log switch, waiting for DBWR to do its thing, and Oracle would
have
> > about 0.2% marketshare as a result.
> >
>
>
> Er.......
>
> Major rule of performance benchmarks for Oracle: make the redo
> logs so large you hardly ever get a log switch and its consequent
> checkpoint. Or at least you get so few you can easily include its
> overhead in the dead periods between "measurement" periods.

Quite so, and my standard advice is precisely this... but that's because the I/O associated with any checkpoint (ie, log switch) is obviously an overhead that must be paid for in performance, not because the database actually grinds to a halt waiting for a checkpoint to finish.

This would seem to be a common misapprehension amongst newbies: "database processing actually stops at a checkpoint, and then resumes when it's finished, doesn't it?". Nope, but it's agreed that making the database do any 'housekeeping' work means there's less grunt for processing your data.

Regards
HJR
>
> <g, d&r>
> Cheers
> Nuno Souto
> nsouto_at_optushome.com.au
Received on Mon Sep 16 2002 - 03:42:38 CDT