Error 14027 When Trying to Configure P2P Topology

I'm beginning to wonder if replication is worth the trouble I've had to spend on it!

Trying to create a simple peer-to-peer replication topology. I can create the initial publication. I can change the property to allow P2P. I can backup the database and I can create a snapshot. When, however, I try to configure the peer-to-peer topology, I get a 14027 error message telling me that ONE table out of 250, or so, "does not exist in the current database." That's pretty troubling given that I can run a select query against that specific table. I would appreciate any information that would help me to understand what's going on or successfully set up replication.

Thanks!

Randy

Hi Randy,

To be honest, I am deeply puzzled by the error that you were getting as well. So let me ask you a few questions which will hopefully narrow down the problem somewhat:

1) Does the name of the table giving you problem contain any special characters ($%#@.\{} etc.) I am especially interested to know if there are any leading or trailing spaces in the table name.

2) Is the table owned by dbo or some other schema?

3) This is a long shot but is the table giving you problems really a table? Can it be an indexed view for example?

4) Is the table giving you trouble a partitioned table? A partitioned view?

For the sake of narrowing down the problem, it may be worthwhile to try setting up a separate (non-p2p) publication containing just the problematic table. It would also be great if you can post the row content corresponding to the problematic table when you execute sp_MShelp_replication_table in the publisher database.

Thanks for the feedback.

-Raymond

|||

Raymond,

Thank you for the response. In answer to your questions:

1. The only non alpha character is an underscore like this: claim_cat.

2. Table is owned by dbo

3. I dropped and recreated the table so it is definitely a table.

4. The table is a lookup with onlyl two columns, claim_cat and claim_cat_desc.

One thing I want to try this morning is setting up the entire replication before doing a full backup from the publisher, then initialize the replications from the backup. Originally, I had backed up the publisher database, restored it to the two subscribers, then created the replication. We'll see.

Thanks again.

Randy

|||

Hi Randy,

I find that your comment about using backups that were taken before the publication was created and the fact that you have dropped and recreated the problematic table to be rather interesting in the sense that it may point us to a potential explanation of your original problem. The current implementation of the p2p wizard makes very strong assumption that all nodes involved have identical schema, so is it possible that the p2p wizard hit the "table not found" error when it is setting up re-publishing from the nodes initialized with the slightly out-dated backup that happened to not contain the problematic table? In any case, it would be great if you can let us know if you have better luck setting up the topology with newer backups.

Hope that helps.

-Raymond