On 8/10/07, Robert Treat <xzilla@users.sourceforge.net> wrote:
> Do you really think people dont know postgresql has replication? I find it
> more that the general use of the term replication implies a built in
> solution, which we do not have.
It seems people on IRC are always asking if we have replication,
though typically it's expressed in language like "high-availability"
and "failover", and typically answered with questions along the lines
of "what do you need exactly?". Shortly thereafter in the ideal case,
the original questioner realizes that the painless, linearly scalable
multi-master replication he or she had in mind doesn't actually exist
in *any* database and he or she didn't really need it anyway. I'd love
to see a list of replication options along with pros and cons of each,
links to relevant documentation, etc.
Robert's original assertion, as I read it, is that people assume we
have a replication option built in to the extent that one can just
twiddle config options to make it work. Such a page as is being
discussed would not only demonstrate that we *do* have replication,
but it would make (more) obvious the fact that many of the neat things
one can do with PostgreSQL require adding on some additional software,
and further that adding modules is generally an easy thing to do. If
we can get more people more comfortable rolling their own system by
gluing packages together, by demonstrating that it's a very common
practice and they won't end up with an unsupportable monster if they
try, I believe we'll make a significant stride on the advocacy front.
- Josh