Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 4 Jun 2010 15:31:58 -0500
From:      Adam Vande More <amvandemore@gmail.com>
To:        Peter Clark <clarkp@mtmary.edu>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: Some GMirror questions.
Message-ID:  <AANLkTilFim5MIH6fwB2IsFAfaKpfpGQbZShPZRB53eiY@mail.gmail.com>
In-Reply-To: <4C093A5B.4040705@mtmary.edu>
References:  <4C093A5B.4040705@mtmary.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
I think you are making this harder than is needs to be.  When in doubt defe=
r
to the Handbook and the man pages.  This also a good page
http://onlamp.com/pub/a/bsd/2005/11/10/FreeBSD_Basics.html

Lastly, I know at one point the 'load' algorithm had some performance
> problems and people were saying to only use 'round-robin'. It seems as
> though some code was committed back in Dec 2009 to fix it's this. Is ther=
e a
> practical rule of thumb to using 'load' vs 'round-robin'?


Use load.  http://www.freebsd.org/cgi/query-pr.cgi?pr=3D113885


> Is this an accurate way to look at it?
> "Round-robin because if you have two disks in a mirror, they=92re both un=
der
> the same 'load' constraints, and it is best to KISS."
>

No. round-robin is a simple algorithm which alternates drive requests.  Als=
o
two identical HD's may be mirrored but they will not really ever be same
state in terms of caching, performance, etc.

--=20
Adam Vande More



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTilFim5MIH6fwB2IsFAfaKpfpGQbZShPZRB53eiY>