Re: [squid-users] CARP setup

Hi Henrik,

Thanks for the response.

Also , should the squid server with the below configuration be made
not to cache anything fetched from the parent caches by specifying
"proxy-only" option ? Is that something that a CARP setup demands? How
do I make a particular request be forwarded to a particular parent
cache ?

Thank you.

-Paras

On Mon, Oct 13, 2008 at 7:18 PM, Henrik Nordstrom
<henrik_at_henriknordstrom.net> wrote:
> Most likely your proxy is not allowed to access the parent.
>
>
> On m�n, 2008-10-13 at 17:42 +0530, Paras Fadte wrote:
>> Hi Henrik,
>>
>> I am getting access control list error when squid makes requests to a
>> parent cache.
>>
>> cache_peer parent1.com parent 11000 0 no-query carp weight=2
>> cache_peer parent2.com parent 12000 0 no-query carp weight=1
>> cache_peer origin.com parent 3000 0 originserver default
>>
>> prefer_direct off
>>
>>
>> parent1 and parent2 are able to contact origin server and fetch the
>> content when a request is made to them directly but when the request
>> is made to the squid server with above configuration which in turn
>> makes a request to parent caches first , I get an access control list
>> error.
>>
>> Any help in this regard would be appreciated and thanks in advance.
>>
>> -Paras
>>
>>
>>
>>
>>
>> On Mon, Oct 13, 2008 at 9:44 AM, Paras Fadte <plfgoa_at_gmail.com> wrote:
>>
>>
>> > Thanks Henrik.
>> >
>> > On Sat, Oct 11, 2008 at 1:39 AM, Henrik Nordstrom
>> > <henrik_at_henriknordstrom.net> wrote:
>> >> On fre, 2008-10-10 at 16:40 +0530, Paras Fadte wrote:
>> >>
>> >>> How do I setup CARP in squid ? Whats a typical setup like ?
>> >>
>> >> Squid only support CARP parents, that is a Squid parenting with a CARP
>> >> like array of Squid servers.
>> >>
>> >> Intra-array CARP routing is not supported (where members of a CARP array
>> >> direct incoming requests to the correct array member within the same
>> >> CARP array).
>> >>
>> >> A typical setup using CARP with Squid is a two layer Squid structure
>> >> with N frontend nodes getting requests from users, and N backend nodes
>> >> doing the bulk of the caching. The frontend nodes route requests to the
>> >> backend nodes using CARP to optimize cache utilization and avoiding
>> >> duplicate storage without the need of spending time on query protocols
>> >> like ICP or HTCP.
>> >>
>> >> Regards
>> >> Henrik
>> >>
>> >
>
Received on Tue Oct 14 2008 - 08:40:41 MDT