Search Postgresql Archives

multimaster (was: Slightly OT.)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



As an aside -- please don't start new topics in old threads.

On Fri, Jun 01, 2007 at 10:42:02AM -0400, gonzales@xxxxxxxxxxxxxx wrote:
> 
> I'm disappointed because SLONY-II has not been released yet to support 
> multi-master replication!  

Well, I wouldn't hold my breath.  Most of the participants in that
project moved on, after concluding either that it wasn't going to
solve their problems, or concluding that it'd cost too much to
develop and support for the likely benefit it would deliver.  As near
as I can tell, development on the project stopped.

The inspiration for the Slony-II project, Postgres-R, has been ported
forward to 8.x series by Markus Schiltknecht.  Last I heard, he was
looking for people to underwrite his work on that project.  So if you
really want those features, the obvious way to do it is to put a
programmer on it, and there happens to be a programmer who has a demo
as his argument that it can be done, and he can do it.  

I think you have to understand, however, that Slony-II or Postgres-R
was not in fact the magic carpet you seem to think it was to be. 
There are some pretty significant limitations to the async
multimaster approach it uses.  To begin with, AFAIK nobody has a
working, production-grade group communication system available for
use by Postgres -- the ones that the prototypes were built on were
pretty hacky, and appeared not to be ready for prime time.  Second,
nobody has come up with any way to make this work with READ COMMITTED
mode, which means you pay a really huge price for the replication. 

My real question in all this is, "What is the problem you are trying
to solve?"  Hot failover using combinations of hardware and software,
and a disk array that can be mounted across two machines, is actually
probably good enough for most cases, assuming it is implemented
correctly (see recent discussion on this topic).  So the availability
piece is mostly solved.  What else do you want?

A

-- 
Andrew Sullivan  | ajs@xxxxxxxxxxxxxxx
When my information changes, I alter my conclusions.  What do you do sir?
		--attr. John Maynard Keynes


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Postgresql Jobs]     [Postgresql Admin]     [Postgresql Performance]     [Linux Clusters]     [PHP Home]     [PHP on Windows]     [Kernel Newbies]     [PHP Classes]     [PHP Books]     [PHP Databases]     [Postgresql & PHP]     [Yosemite]
  Powered by Linux