-
タイトル
Agents are unable to maintain a connection to the Coordinator -
説明
The following warnings are seen in the Change Auditor Agent logs:
"A channel fault has occurred in Coordinator connection"
"Marking Coordinator connection 'CoordinatorServer.domain.com(CONNECTION_ID)' as faulted"
-
原因
One possible cause is an F5 Load Balancer is dropping the connections, leading to the channel fault when the Agent attempts to reconnect on the same connection ID. -
対策
Workaround #1:
If you have F5 Load Balancer, onn the F5 Load Balancer device, increase the TCP idle timeout to a value higher than 5 minutes. As every five minutes a Change Auditor agent tries to establish a connection/communication channel with a Change Auditor coordinator. This will guarantee that on a agent with small amounts of traffic that the agent will still be able to check back in with the Coordinator every 5 minutes, and keep the connection active.
Workaround #2:
The following changes can be made to the Agent Configuration, but it should be noted that for one customer who implemented this, it didn't completely eliminate the connection issues, but it did reduce the frequency from roughly 200 disconnects per hour to roughly 200 disconnects per day:
- Create a new Configuration that will be used by just those servers impacted by the connectivity issues described above. This is done on the Administration Tasks tab, and the Configuration section.
- On the System Settings tab of the new configuration, set the Polling Interval to 60 seconds, the Forwarding Interval to 360 seconds, and the Retry Interval to 600 seconds.
If there is still an issue set the Polling interval set to 60 seconds and the Max events per connection set to 500.