Feature: run ETCD with a separate disk
Every change in control-plane nodes when a cloud provider doesn't support disk resizing can result in etcd data being lost/corrupted. Adding a separate disk for the etcd data would prevent those kinds of disasters and give us more freedom when altering the Terraform control-plane nodes. By default, the etcd database size limit is 2GB ref So even a small disk with hardcoded value would fit into this scenario.
Similar tasks: #641
Exit criteria
- [ ] Kubernetes
control-planenodes have a separate disk for storing etcd database
Theoretically, this could be a small SSD disk. This should leave us with an option to use a cheaper HDD for the OS.
We'll come back to this one probably in August 2023
Closing as a part of a backlog cleanup. Don't hesitate to reopen if needed.