Skip to Content
Dec 20, 2018 at 02:23 AM

Load balancing of RFC calls



I've noticed the following RFC calls during a period of heavy load in the system.

At the time of the problem there is a performance issue on ServerB that cause high CPU utilization and a slow down in processing of RFC calls. We use Round Robin load balancing for all login groups.

I checked these programs using report RSSTAT27


ServerA: 1800 executions
ServerB: 1800 executions


ServerA: 10 executions

ServerB: 1550 executions


ServerA: 8 executions
ServerB: 1500 executions

Please refer to the excellent blog posts below:

What does this mean??

Because RFC_DEST_CONFIRM is executed equally on the servers, does it mean that load is balanced evenly between the servers? In that case why are RFC_PING and RFC_DEST_SHIP executed so many more times on the server where performance was bad? Is it because these are called every times a context is rolled out, and roll outs don't happen so often when performance is good? If this is the case, then does it mean that a drop in RFC performance can trigger an even bigger performance issue because of the extra workload from reestablishing connections when rolling out/in?

Kind regards,