> Hello, all,
>
> I'm back again.  Thanks for the help with table ordering and set
> merging.  Now, I'm stumped by having slony replicate sequences.  Two
> questions:
>
> 1. Is the sequence id space the same as the table id space?  Could I
> have a sequence id #1 coexist with a table id #1?

They're separate "name spaces" (loosely speaking), so you can have a table
#1 as well as sequence #1.

> 2. Is it necessary to explicitly replicate sequences that are created
> by defining SERIAL primary keys?

I'd expect so...

> I ask because my first, naive attempt at adding our sequences has the
> slon daemon failing to sync, with the following error:
>
> ERROR  remoteWorkerThread_1: "select
> "_tii".sequenceSetValue(0,1,'200','1000000000000000'); "
> PGRES_FATAL_ERROR ERROR:  setval: value 1000000000000000 is out of
> bounds for sequence "m_report_links_seq" (1..2147483647)
>
> Is there something obvious I'm missing?

Suggestion: Do a pg_dump of that sequence, and see what it produces.  I
expect that something's funny about the sequence definition.

Perhaps surprisingly, you can dump sequence "public.foo" by the command line:

% pg_dump -t public.foo mydb

It would appear that there's some difference between the sequence
definition on the two database instances.