Re: [squid-users] regarding caching and replication

From: Amos Jeffries <squid3_at_treenet.co.nz>
Date: Sat, 20 Feb 2010 00:00:51 +1300

senthil wrote:
> Thank you very much
>
> Hi henri,
>
> We need to make cached data in squid1 to be cached data in squid2
>
> Is there any possibility .
>
> Note: Squid2 is started if request rate increases beyond certain rate

Ah, that explains a lot about what you are thinking.

In your place I would not bother trying to pre-load squid2. Simply get
it running before load is too high, and let normal client requests fill
the cache according to what the sudden load requires (I expect its
different every time?).

The best you can get is to use a peering relation to provide squid2 with
squid1 as an alternative, maybe faster, source of data than the origin
server.

During regular operation the first client request to squid2 will load
the object into cache and following requests will use it.
I believe this does what you want under the conditions you are working.

You should only encounter problems if squid2 was started too late and
the load was to big. Even then the effect would be better than not
starting it.

Amos

> Henrik Nordström wrote:
>> tor 2010-02-18 klockan 19:28 +0530 skrev senthil:
>>
>>
>>> Is it possible to make cached object in squid1 eg a.gif to be in
>>> squid2 cache as a.gif using squidclient
>>> eg squidclient -h ipofparentsquid(172.16.1.15) -m GET
>>> http://www.example.com/a.gif
>>>
>>
>> No, that makes squidclient fetch the URL from ipofparentsquid.
>>
>> For what you describe you need a peering relation.
>> Regards
>> Henrik
>>
>>
>>
>

-- 
Please be using
   Current Stable Squid 2.7.STABLE8 or 3.0.STABLE24
   Current Beta Squid 3.1.0.16
Received on Fri Feb 19 2010 - 11:00:59 MST

This archive was generated by hypermail 2.2.0 : Fri Feb 19 2010 - 12:00:06 MST