> Well, I'm not getting any bites on this issue, so either I'm the first
> person to try this, or I have bad breath.. ;)
>
> Quick question.  Is the slony_tools.conf file only used at
> initialization and startup of the slony processes?  That is, once I've
> got my replication up and running, is the slony_tools.conf file
> referenced again?  The point is, I'm thinking that I can create multiple
> slony_tools.conf files, one for each database.  Link the appropriate one
> during the startup of slony for that database, then move on to the next
> database once that replication is going for the first one.
>
> It appears to me that the slony_tools.conf file is simply a perl module
> that must be referenced by the code, so I'm assuming I can make this
> work, I just don't want to be hacking a bunch of different scripts.

There is a "start_slon" script that references the slony_tools.conf to
find nodes.  So possibly there's a need to re-reference the .conf file.

However, it is highly preferred to use the more modern (and WAY simpler)
"tools/launch_clusters.sh" for that purpose.  (Indeed, anyone still using
"altperl" tools to run slons should be aware, and you guys know well who
you are! ;-), that I decline to continue to provide support for that!)

At any rate, the usual right answer is that the .conf file is mostly for
constructing the cluster, and once it's running that file ought to be
redundant.