I noticed that slony 1.1 (I have not tried earlier versions) create 2  slon 
processes at startup. The startup log mention a "watchdog process". 

 
My question is : do I need a "external" watchdog script to kill/restart  slon 
if necessary (like the perl watchdog tool packed with slony) or is this  
unecessary with the 1.1 version of Slony ?
 
I use ssh tunneling to let my 2 nodes communicate, and I noticed that if I  
shutdown and restart the ssh tunnel a few seconds later (simulating a network  
failure for the slave node) , I have to restart slon (on the slave node only) 
to  keep the replication going on. Is this normal ?
 
Thanks.
 
C.Miosga
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gborg.postgresql.org/pipermail/slony1-general/attachments/20050812/37ff44da/attachment.html