Geoffrey <lists at serioustechnology.com> writes:
> I'm trying to locate Slony-II documentation, but having a hard time
> doing so.  Is Slony-II real, or is it still in development?

It's Resting In Peace.

A number of not-readily-surmountable problems arose...

- There was a new error condition that would have been inherent to it
  (e.g. - not avoidable without significant redesign to our
  applications) that would have, in a 3 node scenario, led to our
  having ~2/3 of transactions failing, where none would have
  failed with 1 node.  (Issue:  Fighting over a single tuple.)

- Lock management turned out to be more expensive than
  expected

- There isn't a really good, really high performance, BSD-licensed
  group communications system.  (Spread isn't quite faultless...)

Those things seem to have been the essential discouragements.

There were also numerous issues such as the handling of schema updates
where there were still large, nebulous problems to be properly
characterized, let alone resolved.

If Theo Schlossnagle's idea of building a successor to Spread (which
hasn't moved in over a year) turned into something, it might be worth
re-exploring the path.
-- 
"cbbrowne","@","linuxfinances.info"
http://linuxfinances.info/info/lsf.html
"Never insult seven men, when all you're packin' is a six gun"
--- Zane Gray