Correctly terminate orphaned Replication Agent source and target server connections

It is possible for a network, particularly a WAN, to periodically lose connectivity between the Replication Agent and its source or target servers. In this situation, the Replication Agent process may reconnect to the servers and find that the original server-side connection still exists. This causes the Replication Agent to fail to access the servers and log one of the following errors to ctreplagent.log.

Example 1. Source server errors

ERR: Failed to enable replication log position persistence on data source: 780

ERR: Check for two replication agents using the same unique ID (<agentid>)

Example 2. Target server errors

ERR: Failed to lock replication state record on data target. Check for two replication agents using the same unique ID (<agentid>).

ERR: Failed to read replication state on data target: 827

V11

In V11 (and added into this V10.4.1 line), this reconnection logic has been changed allowing the Replication Agent to locate the connection that is using the same unique ID and request it to terminate. When this occurs the Replication Agent now logs the following messages.

Note

This modification requires updating the Replication Agent as well as source and target servers to this new line to obtain corrected support.

Example 3. Source server

INF: Successfully disconnected replication reader using same ID on data source

Example 4. Target server

ERR: Failed to lock replication state record on data target. Check for two replication agents using the same unique ID (RA). ERR: Failed to read replication state on data target: 42 INF: Successfully disconnected replication reader using same ID on data target