Hello, I have 6 squid boxes running all as siblings and they do this work
fine. I am even using delay pools to limit per-user bandwidth over an A
class network (10.0.0.0) and distributed radius authentication for user
log-on onto squid.
The trick is to scatter your users over different caches; i do this from the
proxy.pac (downlodable from httpd server on each cache) sending each user to
a different squid cache based on client's IP with all other squid caches as
hot back-up.
Each squid cache is also able to access internet via one of the available
firewall 1s for high availability.
Seams to be stable and efficent: I save 40% of WAN traffic and users don't
complain at all.
Best Regards and thanks to squid developpers !
Vitaliano Trecca
----- Original Message -----
> On Mon, 20 Dec 1999, Jay Wilson wrote:
>
> > Has anybody setup a second local squid box for redundancy and to share
the
> > load?? I don't see any documentation on the FAQ's for it. What is
involved??
>
> I have not, but would love to see/hear a plan on implementing this as
> well.
>
> But, are you talking about sharing one cache (NFS mounted and
> read/writable by both servers), or separate caches (configuring each
> server to be a 'neighbor' of the other?)
Received on Tue Dec 21 1999 - 05:54:59 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:04 MST