URL:
Reporter RHNID:
Section: -
Language: en-US (English)
Workaround: Point out to students that they need to use common sense when scaling clusters.

Description: Depending on the region the cluster is provisioned in, the student may have only one MachineSet scaled to two (which happens in my eu-central-1 region), or apparently, as the book states, two MachineSet resources each scaled to one machine.

Normally this does not really matter, but when we start cluster scaling exercises, the new number of replicas in the book is for a dual-MachineSet cluster, so it will do nothing to a single-MachineSet cluster that is already scaled to two.

Additionally, when scaling down to one, the single-MachineSet cluster would actually end up with just a single node, which effectively kills HA in router, registry, and all other system pods that require at least two instances to not trigger alerts.