On Sun, 12 Jun 2011 19:44:53 +0530, Deepak Rao wrote:
> Hi,
>
> I am using squid 3.0.21 in reverse proxy mode.
>
> While my clients send requests to our application, the query
> parameters are sent to identify the client, whereas the resource
> provided is a public resource and not specific to a client.
>
> e.g., the urls will be like:
> https://www.public.com/index.pdf?clientID=1234 & my application
> returns with cache-control headers of age greater than 1 day.
>
> I see that squid is not retrieving the resource from its cache when
> the clientID changes. But instead, the request hits my application
> resulting in regeneration of the requested resource.
>
> Is it possible to ignore query parameters for cache retrieval logic?
> Can you help me with the configuration for such a requirement? If not
> possible with 3.0 squid, any higher version supports this?
>
> Thanks,
> Deepak
Sigh. Why do people still do this to themselves?
So...
If the backend produces the desired result when the client ID is
dropped, use a URL redirector to send the clients to a public URL
without an ID attached. Or with a fixed dummy ID that the backend wont
complain about.
You have to do the full 302 HTTP redirect bounce to get the caching
benefits. A "simple" re-write does not cut it.
Amos
Received on Mon Jun 13 2011 - 04:04:56 MDT
This archive was generated by hypermail 2.2.0 : Mon Jun 13 2011 - 12:00:02 MDT