On Thu, Jan 27, 2011 at 3:05 PM, Raymond O'Donnell <rod at iol.ie> wrote:
> which I'll have a go at adapting. I'll have a look at daemontools also,
> as Martijn has suggested.

I vote for daemontools too, we use them in production and it greatly
simplifies some stuff.  We don't define our dataset in it, just our
servers, but if you've got 5 or 6 slon machines up and running, it
makes the /etc/init.d/slony stuff much easier.