So I have:

1. Stop the listeners on both nodes
2. Uninstall slony on both nodes (_slony schema no longer exists after this)
3. Initialize the cluster (re-install on both nodes)
4. Start the slon listeners on both nodes
5. create the set (tables and sequences in a single set)

At this point, there are no errors in the slon log on the
subscriber--all looks normal and good.

6. Subscribe the subscriber to the origin for the single set.

At this point, we get a couple of messages we've not seen before in
all our previous testing and production slony use. For every table in
the set, we get this message:

2008-05-09 16:05:52 EDT DEBUG3 remoteWorkerThread_1: table
"my_schema"."one_of_my_tables" does not require Slony-I serial key

I should point out that I'm not using the special SERIAL option that
tells slony to create a unique key. All my tables have either a PK or
a UC that I explicitly specify in my add table lines like so:

set add table (set id=1, origin=1, id=0010, fully qualified name =
'my_schema.my_table', key = 'my_table_uc1');

Then for the first sequence in the set, which we've given ID = 10, we
get this series of messages in the slon log:

2008-05-09 16:05:52 EDT DEBUG2 remoteWorkerThread_1: all tables for
set 1 found on subscriber
2008-05-09 16:05:52 EDT DEBUG2 remoteWorkerThread_1: copy sequence
"my_schema"."foobar_seq"
2008-05-09 16:05:52 EDT ERROR  remoteWorkerThread_1: "select
"_slony".setAddSequence_int(1, 10, '"my_schema"."foobar_seq"',
'my_schema.foobar_seq')" PGRES_FATAL_ERROR ERROR:  Slony-I:
setAddSequence_int(): sequence ID 10 has already been assigned
2008-05-09 16:05:52 EDT WARN   remoteWorkerThread_1: data copy for set
1 failed - sleep 60 seconds
WARNING:  there is no transaction in progress

No, my create set script does not have more than one sequence with
ID=10 in it, and no, I'm not calling the script twice. I've been
searching online resources, and although I find relevant information,
I have not found a solution. I hope one of you gurus can point me in
the right direction.

Thank you!