I am using primary / slave URIs for ActiveMQ Classic 5.16.0, e.g.:
failover:(tcp://primary:61616,tcp://secondary:61616)?randomize=false&initialReconnectDelay=100
When the primary fails and the backup becomes the primary it takes 10 seconds longer every request to process. Why could be the issue? I am thinking it always tries the first URI and fails and falls back to the second URI that's causing the delay. How to avoid this?
I tried setting updateClusterClients=true thinking might let client know which URI has the become the master but it does not seem to be making any difference.