Jeff Frost a écrit :
> I've got a strange problem with one of my slony slaves that I dropped 
> from replication and resubscribed.  During initial sync it is 
> complaining that sequence id 1 has already been assigned when it's 
> trying to add the first sequence in the set (after syncing all the 
> tables):
>
> http://pgsql.privatepaste.com/a88xkleQyj
>
> Aug 4 18:01:42 proddb1-slave2 postgres[26048]: [707-2] STATEMENT: SELECT
> "_engage_cluster".setAddSequence_int(1, 1, 
> '"public"."hibernate_sequence"','public.hibernate_sequence
> Aug 4 18:01:42 proddb1-slave2 postgres[26048]: [707-3] table')
> Aug 4 18:01:42 proddb1-slave2 slon[26034]: [250-1] 2008-08-04 18:01:42 
> PDT
> ERROR remoteWorkerThread_1: "select 
> "_engage_cluster".setAddSequence_int(1,1,
> Aug 4 18:01:42 proddb1-slave2 slon[26034]: [250-2]
> '"public"."hibernate_sequence"', 'public.hibernate_sequence table')"
> PGRES_FATAL_ERROR ERROR: Slony-I: SET AddSequence_int():
> Aug 4 18:01:42 proddb1-slave2 slon[26034]: [250-3] sequence ID 1 has
> already been assigned
> Aug 4 18:01:42 proddb1-slave2 slon[26034]: [251-1] 2008-08-04 18:01:42 
> PDT
> WARN remoteWorkerThread_1: DATA copy FOR SET 1 failed - sleep 60 seconds
>
> Any idea how to fix this without dropping it from subscription and 
> subscribing it again?  I would expect that no sequences have this id 
> yet since this is the initial sync.  Maybe it's possible I didn't 
> successfully remove the entire slony schema before resubscribing?  
> That would have slonik throw an error correct?
>

Well, something tells me you're facing what we could call a bug, no ?

I had the same problem last week. What version of slony are you runinng ?

Regards,
Stéphane