Re: [SQU] accelerator and icp

From: Ask Bjoern Hansen <ask@dont-contact.us>
Date: Wed, 20 Dec 2000 11:30:06 -0800 (PST)

On Wed, 20 Dec 2000, Henrik Nordstrom wrote:

> You basically can't with a single Squid. The output of the redirector is
> the URL on which Squid caches the object, and there is no second
> redirector step.
>
> What you can do is to chain two Squid's together, where the second uses
> a redirector that removes the unwanted parts of the URL and don't do any
> caching.

Ahh, of course. I just have to figure out where to fit that in with
the (network) fault tolerance stuff I'm trying to get into the
picture too.

Thanks Henrik.

 - ask
 
> Ask Bjoern Hansen wrote:
> >
> > hi,
> >
> > I've setup some squid servers as http accelerators with a simple
> > redirector so http://nameofcluster/foo/blah/server/path will be
> > rewritten to http://server/path?blah. The server is setup to ignore
> > reload requests and with loong expire times. The /blah/ part we can
> > then change to make the squid request the document again. It works
> > great. Squid rocks.
> >
> > How can I make squid not send the ?blah part when it forwards the
> > request? It's not too important, but it would be cleaner.
> >
> > If I enable proxy with accel can I then use ICP to have the squids
> > peer for each other (to provide more fault tolerance)?
> >
> > The servers will be in clusters of 3-7 servers in each geographic
> > location with a number of geographic locations around the world. How
> > good is squid at figuring out which peer to use if I just list a
> > bunch of peers from other clusters (again to make it work even if
> > links between the cluster and the main server goes away).
> >
> > - ask
> >
> > --
> > ask bjoern hansen - <http://ask.netcetera.dk/>
> >
> > --
> > To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
>

-- 
ask bjoern hansen - <http://ask.netcetera.dk/>
more than 70M impressions per day, <http://valueclick.com>
--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Wed Dec 20 2000 - 12:32:10 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:57:02 MST