Hello, first thanks for your reply.
Are you talking about multiple-instance article or about configuring
squid with wccp.
If so I've checked them a few days ago. The problem is that I started
both squid's without problems, each listening on different interfaces
(eth0 and eth1)on port 3128, but only one squid i detected by the router.
The first one that is started connects to the router, no matter which
is, the second does not connect.
And the strange thing is that the router wccp informations shows only
one IP, If I start only Squid2 (only eth1) the router shows the IP of eth0.
May be it has something to do with gre tunnel configuration. :(
Cisco 7200
|
|
Switch - 10/100
| |
| | Linux Box
+---------------------------------------------------+
| Eth0 Eth1 |
| | | |
| +-----------+ +------------+ |
| | | | | |
| | Squid1 | | Squid2 | |
| | | | | |
| +-----------+ +------------+ |
+---------------------------------------------------+
Thanks in advance
Artemis
On 12/14/2010 04:32 PM, Kinkie wrote:
Hi Artemis,
There is a very informative article about this in the squid wiki.
I encourage you to check it out.
On 12/14/10, Artemis BRAJA<a.braja@xxxxxxxx> wrote:
Hello,
I want to achieve load-balancing with to Squid instances on a single
server behind a single IP. All we want is to maximize multi-core CPU
utilisation, because as far as we now Squid is not multi-core aware.
Is any tested way to do with without going to virtualization.
Can it be done using wccp and squid running in transparent mode, so we
can use a Cisco router to load-balance requests?
Thanks
Artemis
