We have setup a Rancher cluster (on prem) as below
1 VM for Rancher server
1 VM for etcd
1 VM for control plane
2 VM for worker nodes
The specifications of each machines are
- CPU : 4
- RAM: 4 GB
- Disk space 100 GB
Apart from the default and system pods that is created with the cluster setup, I am having 1 app with 2 pods running. The pod just echoes the host name.
After running around 100 request, I am noticing disk pressure on one of the worker nodes. Specifically the disk-pressure is reported by the nginx-ingress-controller.
my disks are around 90% empty