Hoi,

I've got a question that I haven't quite been able to find the answer
for in the docs. When writing a slonik script it needs connection
strings so that it can connect to all the different nodes for
administration purposes. One of the network topologies we're looking
at involves some machines behind firewalls and I'm trying to workout
how necessary it is to talk to them for every change. The docs say you
can use SSH tunneling to setup the links. My specific questions are:

1. If you do a MOVE SET command, does slonik connect to every node, or
just the node that is the current origin, or any random node?

2. If you use the moveset() plpgsql function, does it matter which
node it is executed on?

3. Same questions but for FAILOVER and the the failednode() function.

4. Can you tell slonik to automatically run a command before
attempting to setup a connection to a database?

The question I think hinges on what an "event" is, which I haven't
found a good definition for in the documentation (I would have thought
it would be under "concepts"). I think that because MOVE SET is an
event it can be executed anywhere, but because the failover is not an
event it has to be executed on every node individually. Am I right?

Essentially what I'm wondering is, do I need to be able to contact
every node in the cluster directly to do a controlled switchover or
failover?

Thanks in advance,
-- 
Martijn van Oosterhout <kleptog at gmail.com> http://svana.org/kleptog/