Subject: Re: fsync performance hit on 1.6.1
To: enami tsugutomo <email@example.com>
From: Chuck Silvers <firstname.lastname@example.org>
Date: 07/07/2003 09:29:52
ah right, msync() without MS_INVALIDATE on just the range of each write()
should have the same effect as fsync(), but faster. it's been
a while since I looked any of the pgo_put routines, I'm forgetting
them already. :)
On Mon, Jul 07, 2003 at 10:55:05AM +0900, enami tsugutomo wrote:
> Chuck Silvers <email@example.com> writes:
> > if you do continue to access the data after you've written it, then
> > I don't have any good ideas.
> Since the range need to sync each time looks relatively small from
> later post, msync()ing only the range will result tranvasal by range
> rather than list, won't it?