Re: [squid-users] Squid giving TCP_MISS 99% of the time

From: Chris Robertson <crobertson@dont-contact.us>
Date: Fri, 06 Apr 2007 12:33:48 -0800

Adam Membrey wrote:
> Thanks for that Martin.
>
> I've provided links to the logs entries (I apologise for the size but
> it gives you a better idea)
>
> I have also noticed a lot of releases in the store.log
>
> http://users.dodo.com.au/~matildaiglesias/access.log
> http://users.dodo.com.au/~matildaiglesias/store.log
>

Please don't top-post
(http://www.xs4all.nl/~hanb/documents/quotingguide.html). It ruins the
flow of archived messages.

Running the supplied access.log through SCALAR (http://scalar.risk.az/)
I'm seeing a 7% hit ratio on "Images/Graphics & Flash". Out of all
requests resulting in a "200 OK" (4,625) 6% were hits (273). Given
there are only 3 clients and the data given covers ~5 hours with the
bulk of the traffic in one 60 minute time span, that's understandable.
Many of the cacheable requests made will be cached by the browser, and
not requested again (thereby lowering your hit ratio).

The saved traffic ratio is not so good...

     In Traffic: 75.796 MB
    Out Traffic: 76.074 MB
  ------------------------------------
  Saved Traffic: 284.692 KB 0.37 %

The full SCALAR output is available upon request.

Chris
Received on Fri Apr 06 2007 - 14:34:06 MDT

This archive was generated by hypermail pre-2.1.9 : Tue May 01 2007 - 12:00:01 MDT