Date: Sat, 24 Nov 2001 10:45:22 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: Kirk McKusick <mckusick@beastie.mckusick.com> Cc: Sheldon Hearn <sheldonh@starjuice.net>, freebsd-arch@FreeBSD.ORG Subject: Re: Using a larger block size on large filesystems Message-ID: <200111241845.fAOIjM377587@apollo.backplane.com> References: <200111240936.fAO9aXH03886@beastie.mckusick.com>
next in thread | previous in thread | raw e-mail | index | archive | help
:I am of the opinion that we should default to 16K/2K for most
:filesystems today. I believe that the change should be in newfs.
:
: Kirk McKusick
The only thing I worry about is reduced performance when doing
random database accesses, which makes me kinda want to give the
system the capability to do smaller I/O's :-) But apart from that
worry I agree completely. We get fewer indirection levels (64MB
multiplier instead of 16MB per indirection block) , smaller bitmaps
(1/2 the size), and less strain on the clustering code (at least for
sequential I/O). Memory is getting cheap and filesystems are getting
larger, too.
Sheldon, I think you have a go to change the newfs default. Do it!
p.s. side note on the buffer cache: The buffer cache is optimized
for both 1K/8K and 2K/16K, but it is *NOT* optimized for anything
larger. 2K/16K is thus the largest configuration we can use optimally
in regards to the buffer cache.
-Matt
Matthew Dillon
<dillon@backplane.com>
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?200111241845.fAOIjM377587>