If you use ectool commands to query for server information without specifying a particular server in an HA setup, the load balancer will make a random selection out of the cluster and return its data. This behavior and the information it presents may be misleading. Server data commands that are affected by this includes:
You must specify a single server in the cluster with the
--server option with each command. You must run the command once for each server in the cluster if you want information from all of them.
Suppose we have a simple HA setup with two servers foo and bar. If we simply run this command:
the load balancer will select between foo and bar essentially at random. There is no way to tell which server the returned statistics belongs to.
To get the internal timing statistics for the entire two-server cluster, run the following commands: