In my experience, CPU usage only becomes an issue as the server gets low on memory. Here's stats on one I run:
-----------------
Total Cache Size: 56Gb (66% of available disk space)
Total Disk Size: 88Gb
Total Mem usage: 435Mb (85% of available RAM)
Total Mem: 512Mb
CPU Time: 0 - 5 %
CPU Spec: ~700Mhz P3
-----------------
This RedHat 7.0 squid server handles traffic from 4 sibling caches (each with about 150 - 200 concurrent users) and approx. 600 direct users. All the users access the server across an ISDN network, so the load on the server isn't quite as much as it sounds, but it is still a serious amount of traffic.
Basically, the memory usage goes up steadily as the cache grows. You can see above, that a 56Gb cache requires a total of around 435MB RAM. I have also found that different servers often show different characteristics - I always setup a cronjob to send an email report of the disk usage, memory and CPU usage once per day. Once the memory used by the Squid processes reaches about 90 - 95% of available RAM, you will see a massive drop in performance, and 100% CPU usage! Once a server gets close to this point, I use the cache size 'high' and 'low water' marks in squid.conf to keep the cache size just right to keep Squid's memory usage around 85%
I think that if you get a 1Ghz P3 (which isn't expensive these days) you'll need a massive throughput to cause it problems. Just remember to get plenty of memory!
Aaron
-----Original Message-----
From: Colin Campbell [mailto:sgcccdc@citec.qld.gov.au]
Sent: 23 October 2002 06:56
To: Jaime Elizaga Jr.
Cc: squid-users@squid-cache.org
Subject: Re: [squid-users] suggestions for high-end dual-CPU config on
Linux?(+tuning of current conf)
Hi,
On Wed, 23 Oct 2002, Jaime Elizaga Jr. wrote:
>
> ----- Original Message -----
> From: "Colin Campbell" <sgcccdc@citec.qld.gov.au>
>
> > - dual CPU probably won't get you anything as squid is single-threaded
>
> So it would be better off for squid to run on a single FAST CPU than a dual,
> not so fact CPU?
Depends. On some OSes you get "worker" threads for doing diskio that can
make use of the extra cpu. Squid itself will run on one cpu and one only.
I'd be pretty sure putting two "slow" cpus in would be worse than one
"fast" one, but it depends on the difference between "fast" and "slow".
You'd probably get more benefit out of faster disks than two cpus.
I sit here waiting to be shot down :-)
Colin
-- Colin Campbell Unix Support/Postmaster/Hostmaster CITEC +61 7 3227 6334Received on Wed Oct 23 2002 - 03:45:59 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:10:52 MST