Hello!
Interesting question, but I've never heard of such a (supported) function.
The method that definitely works, but should of course only be carried out during a maintenance window and only if there are no problems during this step, would be to simply run “clusterXL_admin down” on the other two nodes for the duration of the test and then “clusterXL_admin up” again afterwards. If the remaining node then predictably crashes under the load, this is obviously not a particularly good idea.
I'm curious to hear the ideas of the other mates.
and now to something completely different - CCVS, CCAS, CCTE, CCCS, CCSM elite