Azért megvan ennek a másik oldala is.. Például kubernetes-novice channelen (https://kubernetes.slack.com) pont tegnap dobtak be egy hasonlót:
"One of the developers deploying apps on my kubernetes cluster decided he was too special to provision storage for his pods. He ended up doing a lot of reads/writes to disk which ends up using the docker container's storage. Well for us that happens to be the root volume (We use TwinBlade's for this cluster. Drive 1 is the root drive, drive 2 is a 2TB SSD thats part of a storage cluster). Super high I/O on the root drive ended up slowing the system down, slowing down Docker responding to Kubelet health checks causing Kubernetes to mark the node down. The issue cascaded though, because the trouble-maker pod in question would then get moved to another "Ready" node by the scheduler."
____________________________________
Az embert 2 éven át arra tanítják hogyan álljon meg a 2 lábán, és hogyan beszéljen... Aztán azt mondják neki: -"Ülj le és kuss legyen!"..