I think you need to set each varnish server as a backend, and then hash the URL to a backend (that is a varnish server). If the backend matches the current varnish server serving the request then you use the real backend (origin). I am not sure the details of the setup, but Artur from Fastly said they do something similar in their setup maybe one of the Fastly guys or someone else can comment on how this is done ?
Curious myself as to how to make this work On Wed, Aug 21, 2013 at 12:07 PM, <[email protected]> wrote: > An older blog post on the Varnish site says this: > > "... you can emulate ICP-like behaviour by writing VCL code to check your > neighouring cache for the request that you are looking for. Setups like > these have been discussed on > varnish-misc<http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc/>several > times and implementing it should be fairly simple." > > I must be searching the wrong terms, because I can't find any descriptions > of this. > > It looks to me like the setup would be little more than setting the peer > as a backend and then setting req.hash_ignore_busy on one of the servers to > avoid race conditions, as described on the wiki. Is there anything I'm > missing, or does someone have a reference on how to set this up? Any help > is appreciated. > > _______________________________________________ > varnish-misc mailing list > [email protected] > https://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc >
_______________________________________________ varnish-misc mailing list [email protected] https://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc
