The types of situations that require a full rewrite are the ones where a critical tool was used that was completely wrong for the job, or where key assumptions were made (possibly many of them) that later turned out wrong, or where complex business logic has become so hopelessly spread out through the edges of the application that every new feature requires massive cut and paste.
How does one go about setting up a multi-master replication scheme, also having slaves in the mix? There are several articles online pertaining to how to set up 'multi-master' replication. It seems though, that they are tailored for 'circular' or 'ring' replication. One of my favorites is Giuseppe Maxia's article at ONLamp which I followed in setting up our replication scheme.
I recently reported on my good results with migrating an old server from RH7 to Debian Sarge, beginning with the network-install CD. Responding to my comments that an install of a Sarge desktop on my old Thinkpad left me less than impressed, several readers admonished me for not installing Etch.
I've now done an Etch install on the same Thinkpad I recently used for evaluating seven single-CD distros, and this article explains what I did and how it worked out.
[This guy] created a file called my-custom-backup.conf based on this and I think others might appreciate having it, since it fixes 2 bugs or adds 2 features (depending on how you look at it) to os2006 - it backs up the application repository list and all archived email ...