On 1/18/2010 11:51 AM, Andrew Sullivan wrote:
> On Mon, Jan 18, 2010 at 07:07:30PM +0300, Nickolay wrote:
>> PostgreSQL 8.3.5
> 
> I note an awful lot of fixes after that in the 8.3.x branch.  Are you
> quite sure you don't have a data-loss problem in Postgres?  I barely
> scanned the release notes & nothing leaped out at me, but that would
> be the first thing I'd fix.  Note that a response of, "But this only
> happens when I'm doing Slony things," is not going to get a
> sympathetic hearing: Slony is already working at the limits of
> Postgres functionality, and it will exercise parts of the system that
> you won't.  If there's a serious problem fixed by the four subsequent
> minor 8.3 releases, I'll bet a pretty good lunch Slony will tickle it.
> 
> (This isn't to say that _is_ the problem, but I think you should rule
> it out first.)

As a general rule of thumb, nothing (not even slony) is supposed to be 
even able to break transactional integrity. It is PostgreSQL's job to 
deny that under all circumstances.


Jan

-- 
Anyone who trades liberty for security deserves neither
liberty nor security. -- Benjamin Franklin