Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 14 Dec 2006 23:49:40 +0100 (CET)
From:      Wojciech Puchar <wojtek@tensor.3miasto.net>
To:        Dieter <freebsd@sopwith.solgatos.com>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: disk I/O tuning parameters
Message-ID:  <20061214234546.C32744@chylonia.3miasto.net>
In-Reply-To: <200612132215.WAA11949@sopwith.solgatos.com>
References:  <200612132215.WAA11949@sopwith.solgatos.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> Is there a different, more detailed, description of
> its effects (and/or similar tuning parameters) than found in
> http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/configtuning-disk.html
>
> Is there a way to limit the runningspace, bufspace, or similar parameters
> on a per disk, per process, or per file basis rather than system wide?
> I haven't been able to find anything.
>
> I need a way to protect the disk I/O bandwidth of one process from
> other processes.  Having its own disk, and running at rtprio is not

IMHO no, while it could be useful.

priority schedules only CPU usage.


BTW - if you are processing huge files often consider editing 
/usr/src/sys/sys/param.h

and change

#define MAXPHYS         (128 * 1024)   /* max raw I/O transfer size */

to

#define MAXPHYS         (1024 * 1024)   /* max raw I/O transfer size */


it won't make disk I/O scheduling better (even a bit worse) but will 
GREATLY speed up huge file I/O. bump hi and lorunning space at least twice 
the default (4 and 2MB) maybe more if you have many concurrent disks.

i'm using that settings on every FreeBSD machine, no problems.


when making new filesystem set -a option right when doing newfs (read 
man).


PS. it's not answer for your question but it may help either



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061214234546.C32744>