On Fri, 3 Dec 1999 18:00:32 +0100 , [EMAIL PROTECTED] writes:
> On Fri, Dec 03, 1999 at 10:44:51AM -0500, Mark Hoffman wrote:
> > I've got the same problem. Forgive me if the answer is documented
> > somewhere, but does anyone have a solution? I've put in all the
> > patches for high volume servers.
>
> My solution was to remove the 'sync' option from the queue mount. It
> now races it's ass off, I benchmarked it at well over 1 million
> distinct messages a day.
>
> Still waiting for a reliable filesystem that doesn't come to a
> grinding halt when mounted sync. Under Linux, that is.
Well, under FreeBSD, 'sync' only does sync metadata,
which is a significant performance improvement over
sync data+metadata. Also, the softupdates option
(which is allegedly unsafe, but I haven't seen anyone
say why) will let you handle a bit more volume still.
>From experience, softupdates gives you almost as
much improvement over (FreeBSD) sync as async gives
you over softupdates.
Aside from that, I guess you're stuck waiting for
XFS or ext3fs journalling, whichever comes first.
--
Chris Mikkelson | It was mentioned on CNN that the prime number
[EMAIL PROTECTED] | discovered recently is four times bigger than
the previous record. -- unknown
