This message was deleted.

Is this due to using the same machine_config_v2 resource for both pools,

Yes. Also keep in mind that a

machine_config

is immutable - not necessarily reusable, so if your cluster is referencing a single

machine_config

, any changes will have a cluster-wide impact (probably not what you want). There should be a 1:1 relationship between a

machine_config

&

machine_pool

, which you can use a

for_each

to accomplish. As an example: https://github.com/frank-at-suse/vsphere_HA_autoscale_cluster/blob/master/cluster.tf. This behavior is related to CAPI, which RKE2 uses, and is not Rancher-specific.

Thank you for your reply. Then my assumption was just not right 🙂, as referencing a machine_config by kind an name suggests that is is reusable. And also thank you for your example.