On Fri, Jun 30, 2006 at 11:39:04AM -0400, Tom Lane wrote:
> Scott Marlowe <smarlowe(at)g2switchworks(dot)com> writes:
> > On Fri, 2006-06-30 at 08:17, Jason McManus wrote:
> >> * Replication support still rudimentary.
>
> > Hmmmm. I think that's an overly simplistic evaluation. The slony
> > replication engine is actually VERY advanced, but the administrative
> > tools consist mostly of "your brain". hehe. That said, once you've
> > learned how to drive it, it's quite amazing. Keep in mind, slony can be
> > applied to a living database while it's running, and can run between
> > different major versions of postgresql. That's a pretty advanced
> > feature. Plus, if the replication daemons die (kill -9ed or whatever)
> > you can restart replication and slony will come right back where it was
> > and catch up.
>
> It might be worth pointing out that mysql's replication falls over
> if you so much as look at it crosseyed. I have not had to use it
> for production purposes, but I can tell you that the mysql replication
> regression tests fail ... irreproducibly of course ... almost one time
> in two in Red Hat's build environment. I've been able to trace a few of
> these failures to quirks of the build environment, like trying to build
> x86 and x86_64 at the same time in different chroots of the same machine
> (must take care not to use same TCP port numbers for tests), but it
> still seems flaky as hell.
I attended a talk about MySQL and High Availability once and was pretty
unimpressed. Lots of 'now you take the database down and copy files
around' and the like. Nothing remotely close to the abilities of Slony.
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461