Restart of Edge components

When stopping Edge, as the page says the order is important. So if I make a config change in Cassandra e.g to the default.properties file do I have to bring down all the subsequent nodes before that. OR can I just bounce Cassandra? This goes for any nodes.

If I want to bring down the router? Do I have to bring down the UI & MS first?

0 1 213
1 REPLY 1

Components like RMP (router and message processors) and management nodes can be restarted independently of each other. For zero downtime of your runtime traffic, you may have a load balancer (LB)in front of your RMP nodes, so that you can reroute traffic to the RMP node that is up while taking down the other for restarts. Once the RMP is restarted you add it back into your LB.

Management server server restarts shouldn't affect runtime traffic unless your APIs are making management calls as a part of a callout (which is typically not recommended). So restart of management server would simply take down the ability to run management calls or for the UI to run the management calls.

Datastore nodes we have to be more careful in cases of runtime traffic. So best to get your topology to understand what you can do here.