True indeed. complexity does bring in the failure as it has for me. I've reverted back to one master and one backup and it works like a charm. I guess the electricity we have to pay for on the backup is the price we pay for high availability. Perhaps I could make the backup useful for DNS caching/memcaching among other things.
Thanks and I do believe this thread to be resolved for all intensive purposes, Vivek On Wed, Nov 19, 2008 at 12:51 AM, Rogier Krieger <[EMAIL PROTECTED]> wrote: > Sent off-list; not intended as a flame but to save you some pain. > > On Tue, Nov 18, 2008 at 21:21, Vivek Ayer <[EMAIL PROTECTED]> wrote: >> I may actually end up just turning off load balancing on the router >> for now and just leave it on the web servers. > > That changes the [routers'] job description to routing and proper > failover. If you have a test network, you can try working out the > kinks and a working setup that includes load balancing again (if you > need it). > > >> Is there a clever cron script I could write to manually change the master >> on a day to day basis? [...] It would be nice if the routers took turns >> every day rather than every few seconds/minutes. > > Your primary router is there to do work; your backup is there to take > over in case of trouble or when you want to take the primary down > (e.g. for maintenance). Why risk it and for what tangible benefit? > > The more complexity you introduce, the more things are likely to fail. > (Courtesy Nick Holland). > > Greets, > > Rogier > > -- > If you don't know where you're going, any road will get you there.