Books on writing distributed systems

[Simon Wistow]

Is there any decent literature on writing systems that keep data in sync 
in a master-master/live fail over kind of way. I seem to have 3 projects 
on that all have that requirements in that area. 

To be clear - I'm aware that there *is* literature out there on this 
stuff, I'm just looking for a recommendation. At the moment I'm leaning 
towards Tannenbaum's book because, well, he rocks.

Additionally, are there any libraries out there that generically do this 
sort of thing i.e have a binary blob of data that's kept in sync between 
n>1 machines?