intermediate · ~8 min

The right hardware

The cache is slow. Ops explains: only hackademy-worker2 has the fast SSD — the other worker has spinning disks — and the scheduler put the cache wherever it liked.

Nodes carry labels like any object. A pod's nodeSelector lists labels a node must have before the scheduler will place the pod there. kubectl label node NAME key=value adds one; kubectl get nodes -L key shows it as a column.

Objectives

Label hackademy-worker2 disk=ssd (and only that node)
Run both cache pods on SSD nodes only
Bonus: See the label with kubectl get nodes -L disk

Best on a computer with a keyboard. It works on a phone, with shortcuts under the terminal.