On Thu, 21 Dec 2006, Alvaro Lopez Ortega wrote: > Stefan de Konink wrote, On 21/12/06 21:29: > > >> Pretty useful, don't you think? :-) > > > > Yes yes :) So what would a user do to prevent single point of failure and > > single bandwidth aggregation? If a heartbeat was added to this cute > > thingie... N-servers can serve a M-size cluster for X requests. > > > > Then you are talking about DNS updating meganism ofcourse or some (smart) > > routing. > > No, it doesn't need any sort of DNS tweaking. The client requests go > through Cherokee to a destination server (that usually are placed in > the intranet), and the outgoing replies goes through Cherokee as > well. All the traffic seems to be generated by the public Cherokee > server, even if it is actually generated by server in the intranet.
If the mirror server is down. Everything is down. It would be nice if it was possible to have some round-robin system in there. That actually takes over if one of the mirrors fail... > > So in a simple way: is it possible to have some 'callback' feature > > or status interface so it is possible to link it to a third party > > management tool? > > I'm not sure about what you asking. > > Cherokee includes a remote administration handler called "admin", > maybe that's what you are looking for. I'm was thinking about something like 'if the handler fails then execute command'. Stefan de Konink _______________________________________________ Cherokee mailing list [email protected] http://cherokee-project.com/cgi-bin/mailman/listinfo/cherokee
