You can use a director to have the other varnish as a fallback of your real backends, and maybe play with the weights and retries to keep you varnishes warm.
Best Regards, Dridi On Mon, Feb 18, 2013 at 2:46 PM, Rafał Radecki <[email protected]> wrote: > Interesting, what if one of the servers goes down? Should the backends > be defined in order? For example: > - varnish1 asks varnish2 > - if varnish2 does not answer it asks backend directly > ? > > 2013/2/18 Dridi Boukelmoune <[email protected]>: >> Hi, >> >> Does this help ? >> https://www.varnish-cache.org/trac/wiki/VCLExampleHashIgnoreBusy >> >> Best Regards, >> Dridi >> >> On Mon, Feb 18, 2013 at 1:37 PM, Rafał Radecki <[email protected]> >> wrote: >>> Hi all. >>> >>> I have a two node pacemaker cluster. If primary node goes down then >>> secondary node takes its role. >>> I use varnish on both nodes. When I switch to the secondary node the >>> cache needs to be regenerated (as expected). Is there a way to >>> replicate varnish cache between two servers? Something like conntrackd >>> for iptables? >>> >>> Best regards, >>> Rafal Radecki. >>> >>> _______________________________________________ >>> 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
