We had a hung JVM today and the gateway failed over as expected. When the backup node was active, however, no clients could connect to it. We have followed the guidelines given to us by support that stated to turn off HTTP Auto-detect and to specify both gateways in the HTTP address section.
The client launcher log has
ERROR [SplashFrame ] [11:10:43,471]: Gateway is running, but is not active. Status is: Cold
ERROR [SplashFrame ] [11:10:43,471]: Backup Gateway is running, but is not active. Status is: Cold
But this message is inconsistent, since both the master and backup gateways show the backup gateway as active.
We tested this scenario in 7.7.6 and it worked. Is there new guidance for 7.8.2 as to how redundancy should be configured?