Hi all, Up to now, i'm been running "standalone" varnish on 2 servers (for redundoncy).
I've been asked to have those 2 varnish "work together" : if there is miss on server1, server1 asks server2 (in case server2 has the object in cache), before asking the backend Do any of you have any comments/thoughts about this ? (it does make sense..i think) This is doable using Directors in vcl (a failover algorithm would be optimal, rather than round-robin or random) Any other solutions ? Olivier _______________________________________________ varnish-misc mailing list varnish-misc@projects.linpro.no http://projects.linpro.no/mailman/listinfo/varnish-misc