Thanks for the reply amos.
I'm a total newbie to this kind of configuration, could you give me an
example of using cache_peer sibling statement?
also what options do i have to implement the LB before it goes to the
chosen squid server?? have in mind again i'm a total newbie regarding
this type of secenario...
Thank you!
On Sun, Dec 19, 2010 at 4:52 AM, Amos Jeffries <squid3_at_treenet.co.nz> wrote:
> On 19/12/10 11:51, N3O wrote:
>>
>> Hi
>> Does anyone know how to implement a two node squid reverse proxy??
>> My idea is to have 2 squid servers working as reverse proxy to an internal
>> apache web server. The two node should do some kind of load balance
>> between them.
>
> What sort of capabilities do you have around the network to do that LB?
> LB must be done before the request enters into the worker squid. (could be a
> hardware LB, some software scripts, routing rules or another proxy).
>
> For surety once a request enters either of the worker squid it may as well
> be processed by that one. The only benefit of sibling links is when the data
> is cached in the sibling for fast retrieval. Double-handling is a problem.
>
>> So a simple design would be:
>> internet
>> |
>> |
>> LB
>> squid1 -- squid2
>> \ /
>> \ /
>> \ /
>> apache
>> Which would be the best idea to implement this scenario?
>> Thanks!
>
> There is nothing special involved.
>
> 1) Setup each node as a reverse-proxy separately with whatever handling is
> appropriate for your needs.
> 2) Test that works.
>
> 3) Add the cache_peer sibling link between.
> 4) Test that works.
>
> 5) setup the LB to pass requests between them with whatever balance you
> like.
>
> Amos
> --
> Please be using
> Current Stable Squid 2.7.STABLE9 or 3.1.9
> Beta testers wanted for 3.2.0.3
>
Received on Sun Dec 19 2010 - 16:05:31 MST
This archive was generated by hypermail 2.2.0 : Tue Dec 21 2010 - 12:00:03 MST