So testing switchover after upgrading to 1.2 and 8.4.3 (well some are
still on 8.3.4).

And I'm getting this error when trying to switchover.  I get this
error when I tried to drop set 3 (thought about creating another set
to see what the issue is).

I have 9 tables, and 7 of the 9 seem to work, but slon seems to be
missing, not replicating and in fact citing that the table with the id
can't be found, yet I find it just fine.

Even if there is no clean way to "fix" this, is there a hack to drop
this set in order to try to recreate the set for testing? Odd that
it's only these 2 "ancient" tables are causing grief.

Ideas?

Thanks
Tory


-bash-3.1$ ./switchover_to_node2.sh
<stdin>:22: Moved Set 1
<stdin>:29: Moved Set 2
<stdin>:34: PGRES_FATAL_ERROR select "_cls".moveSet(3, 2);  - ERROR:
Slony-I: alterTableRestore(): Table with id 46 not found
CONTEXT:  SQL statement "SELECT  "_cls".alterTableRestore( $1 )"
PL/pgSQL function "moveset_int" line 33 at PERFORM
SQL statement "SELECT  "_cls".moveSet_int( $1 ,  $2 ,  $3 , 0)"
PL/pgSQL function "moveset" line 61 at PERFORM
Switchover complete

Tables are in there and slon schema knows about them.

   46 |   24509700 | profile_fields          | adm   |       3 |
profile_fields_pkey              | t           | profile_
fields table
     47 |   24509710 | profile_values          | adm  |       3 |
profile_values_pkey              | t           | profile_
values table

Yes the switchover failed and I just removed the lock on set 3, since
it didn't actually move.