Bill Moran wrote:
> I've been putting this together over the last couple weeks for
> the reasons listed on the HTML page:
> http://people.collaborativefusion.com/~wmoran/PostgreSQL/slony_switchover.html
> 
> I'm interested to hear how useful this is to others, and of course
> suggestions for improvement are welcome.

This looks good Bill.  I like how you build the preamble from the
database instead of relying on slonik config files.  We currently rely
on slonik_move_set to do an origin "switchover", which requires that we
maintain an accurate preamble file on all our database nodes.  Keeping
the preamble and other slon* configs up to date is cumbersome (although
we use puppet for this, which helps).

Some questions for you:

You put a 'wait for event' command after 'lock set' *and* 'move set'.
slonik_move_set only waits after 'lock set'.  Is this extra wait
necessary?  If so, should it be fixed in slonik_move_set?  I'm using
this version:
   slonik_move_set.pl,v 1.1.4.1 2006-10-27 17:54:21


you state here
http://people.collaborativefusion.com/~wmoran/PostgreSQL/slony_switchover.html that...
> Our network topology causes connection paths to vary depending on
> where you're running the slonik script from, thus each node would need
> its own slonik script that was different than other nodes

So there's more than one way to connect to a given node?  Can you
explain this further?  What kind of network topology do you have that
necessitates having connection paths that vary?  (I obviously need to
read more slony documentation)


Thanks for the script (-:

Dane
-- 
Dane Miller
Systems Administrator
Greatschools.net http://www.greatschools.net