Date: Fri, 7 Dec 2001 14:57:45 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: Sheldon Hearn <sheldonh@starjuice.net> Cc: Kirk McKusick <mckusick@beastie.mckusick.com>, freebsd-arch@FreeBSD.ORG Subject: Re: Using a larger block size on large filesystems Message-ID: <200112072257.fB7MvjE95211@apollo.backplane.com> References: <31807.1007732134@axl.seasidesoftware.co.za>
next in thread | previous in thread | raw e-mail | index | archive | help
:Hi Matt,
:
:I already have patches for newfs, sysinstall and the tuning(7) manual
:page which simply make 16384/2048 the default.
:
:However, I remember people expressing concerns with using these parameters
:for very small filesystems.
:
:I can't find any negative performance impact for very small filesystems.
:
:The only other thing I can think of is what obrien suggested. He told
:me that it might be that people are wary of a filesystem that contains
:only a single cylinder group, as this means you only have one
:superblock.
:
:Is this really something to worry about? Is there some other reason why
:we shouldn't just make 16384/2048 the blanket default?
:
:Ciao,
:Sheldon.
I think you've hit the nail on the head. I think we do want at least
two cylinder groups, but beyond that I think even the smallest hard drive
will do just fine with 16K/2K. Floppies and other types of media are
already hardwired in /etc/disktab so we shouldn't have to worry about
them.
In considering Kirk's posting in regards to the 512/4K -> 1K/8K debate,
I think we are in exactly the same situation here and should just
maintain a stiff upper lip and change the default to 2K/16K. People
with 'special' requirements are smart enough to manage the parameters
themselves. 2K/16K is a good default for more of our users then 1K/8K
was.
-Matt
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200112072257.fB7MvjE95211>