Re: 100-110 response/s Polygraph performance from Squid 2.2 using Solaris 2.7 and "RAMdisk" on 270 MHz Ultra5

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Sat, 11 Dec 1999 02:43:26 +0100

Peter Jenny wrote:

> I thought by eliminating hard disks Squid would perform much faster, but
> I'm only seeing ~110 responses/second from my lab test configuration:

If you want a baseline of what the CPU is capable of, try

no_cache deny all

> I've seen some messages in the (excellent and still up) mailing list archive
> at http://www.progressive-comp.com/Lists/?l=squid-users&r=1&w=2
> to the effect that UFS is the problem; I guess even when the file system
> physically resides in RAM it's still a major performance problem?

Plain Sun UFS implementation is perhaps not the fastest filesystem
available. However, tmpfs is not UFS and have quite different
performance characteristics and oddities than UFS.

A guess is that tmpfs is optimized at holding a few transient files of
varying size, not as a large semi-permanent storage containing many
files.

If you want to find the UFS performance if the disk was unlimited in
speed then have an equal amount of ram as you have disk, mount the
filesystem with noatime and run fastfs on it.

--
Henrik Nordstrom
Squid hacker
Received on Fri Dec 10 1999 - 18:55:33 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:49:50 MST