Hello good slony-1 people--

i seem to have gotten a slony replication set into a pretty bad state
with some experimentation.  could you help me get out of it?

the cluster is named foo. I have 2 sets and 4 nodes in the cluster,
nodes 1 and 2 belong to set 1 and nodes 3 and 4 belong in set 2.

node 4 is subscribed to node 3, and node 2 is subscribed to node 1.


Somehow, node 4 knows about itself, but none of the other nodes do.

if i try the slonik command:

  store node(id = 4, event node = 1, comment = 'node 4');

I get the following response:

<stdin>:8: Error: namespace "_foo" already exists in database of node 4


This is true, because the node was set up before, but apparently wasn't
properly registered with the other nodes (or i broke the registration
somehow).

If i try the following slonik command:

  drop node (id = 4, event node = 1);

Then i get:

<stdin>:6: PGRES_FATAL_ERROR select "_foo".dropNode(4);  - ERROR:
Slony-I: unknown node ID 4
<stdin>:8: Failed to drop node 4 from cluster


Is there a way to recover from this state without dropping the entire
4th database entirely and starting over?  any ideas how i got into this
situation, and how to avoid something like this in the future?

I'm happy to provide more details if anyone wants them.

	--dkg

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1030 bytes
Desc: OpenPGP digital signature
Url : http://lists.slony.info/pipermail/slony1-general/attachments/20110713/77e131d1/attachment.pgp