Er, no, I suppose I should have said that I get the 15 second warning, 
then the 30 second warning, and then as many 60 second warnings as I 
want before I give up and kill the slon processes. AFAIK there are no 
open transactions, and yet slony thinks there is. It seems like the 
error is that the slave slon thinks there is no remote worker 
thread.... how could I validate if that was true?

On Aug 30, 2004, at 4:48 AM, Jan Wieck wrote:

> On 8/30/2004 12:49 AM, Ben wrote:
>
>> Both master and slave databases are brand new.... maybe 25 seconds 
>> old by the time I start slon. :) So no stale transactions laying 
>> around. In fact, they haven't seen any connections at all, save from 
>> the slony scripts.
>> Might this be because my postgres version is a little behind the 
>> times?
>
> After all it is a warning, not an error. Wait for 15 seconds, don't 
> slam on the red panic button, and the subscriber slon will try it 
> again. If there are no others that slonik's and slon's causing open 
> transactions as you say, that'll do just fine.