Details
-
Type:
Bug
-
Status:
Open
-
Priority:
Major
-
Resolution: Unresolved
-
Affects Version/s: 1.7.1
-
Fix Version/s: None
-
Component/s: Core
-
Labels:None
-
Environment:RHEL 5.0
-
Number of attachments :
Description
We have been working to create a framework around SymmetricDS that would allow users to configure it using minimal steps. As part of this framework, we provide a way for users to express their interests in tables under various schemas.
It would be apparent to you by now that we DO NOT create triggers on tables as part of our initial configuration. The slave node sends its requirements (tables) to its master. The master then creates triggers on the required tables. In essence, we do a trigger on-demand.
The issue that we are facing with this kind of architecture is that after creating the triggers and calling the 'syncTriggers' function on the 'Node' Mbean (on the master); the triggers are not sync'ed onto the slave. This effectively means that there are no parameters (target schema, target table etc.) available while loading the data on the slave. As a result replication does not occur.
While trying to do some sanity testing, we stopped the slave SymmetricDS instance, removed the registration information of the slave on the master and restarted slave. This time the triggers were successfully sync'ed and replication started.
Note: We use SymmetricDS 1.7.1. We use PostgreSQL database 8.3.7. We have 'auto.registration' set to true on master and 'dataloader.lookup.target.schema' set to true on the slave. We do cross-schema replication. But in our scenario replication was between same schema names.