I'm getting the error Unable to join cluster: System 'systemname' already has a record in the active systems table and appears to be active. When trying to cluster my MFT nodes. What would be causing this?
Here are the three most common issues:
- Most often, this issue occurs when you have the same system named defined in the cluster.xml for two or more MFT nodes. To fix this, rename your cluster.xml to a unique name. Best practice would be to name it something easily identifiable, such as a DNS name.
- When there are one or more firewalls between the MFT systems, the nodes in the cluster cannot communicate if the cluster bind port is not open on both MFT systems. When that occurs, the GoAnywhere MFT nodes fail to start. To fix this, make sure each node can reach the other node on the cluster bind port as defined in the cluster.xml file. A good test for this is to use the telnet command from one node to the other cluster node's bind port.
- The system name is still in the dpa_active_system table inside the database. This can occur when the system is unexpectedly shut down and the node is not able to send the proper shut down commands to the database. For example, a hard reboot of the MFT server. To fix this, you will need to work with your database administrator to remove the row for the inactive system
e. [email protected]
w. GoAnywhere.com | HelpSystems.com