Date: Sat, 30 Jul 2005 15:31:40 -0700 From: Julian Elischer <julian@elischer.org> To: Poul-Henning Kamp <phk@haven.freebsd.dk> Cc: FreeBSD Current <freebsd-current@freebsd.org>, Brian Candler <B.Candler@pobox.com> Subject: Re: Apparent strange disk behaviour in 6.0 Message-ID: <42EBFFCC.5010603@elischer.org> In-Reply-To: <5502.1122754726@phk.freebsd.dk> References: <5502.1122754726@phk.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp wrote: > In message <42EBD958.6040402@elischer.org>, Julian Elischer writes: > > >>>If you know your queue-theory, you also know why busy% is >>>a pointless measurement: It represents the amount of time >>>where the queue is non-empty. It doesn't say anything about >>>how quickly the queue drains or fills. >> >>exactly.. I'm trying to work out why teh read and write queues are empty for so >>much time in a transaction that SHOULD be disk bound.... > > > I am very confident that the disk statistics collected in GEOM don't lie: > your disks are idle because nobody submits I/O requests. > > Look at your scheduler... I plan on doing just that.. I didn't say it was a problem in geom.. I just said that it is worth investigating why there is no cannel saturation when in theory that should be teh bottleneck. I have noticed this before.. 5 & 6 sometimes just don't seem to be "trying". they often have idle time when I would expect none.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42EBFFCC.5010603>