How to reset the Kubernetes cluster with Torus

94 views

Skip to first unread message

Christopher Schmidt

unread,

Jul 4, 2016, 1:41:09 AM7/4/16

to CoreOS User

Hi, I am using Torus (v0.1.1) as a volume plugin in a Kubernetes (v1.2.3) Cluster with Postgres and Mongo volumes.

After creating my Pods, I get some errors like "status: Failure, reason: Unit torus-dev-nbd0.service already exists." etc.

However, I am unable to somehow reset the cluster, to do the tests again.

What do I have to do to reset the Torus setup (except recreating the whole cluster)?

regards Christopher

(k8s-node.j2 is my Ansible node template, that I use to create the cluster)

barak.m...@coreos.com

unread,

Jul 6, 2016, 2:03:20 PM7/6/16

to CoreOS User

It's probably being held open by not closing correctly (this is a prototype after all :) )

Simplest thing is to just kill the service `systemctl kill torus-dev-nbd0.service` on the affected machine. Once it's gone, it'll be back to the started state.

This will get a lot better (and faster) in a coming release, where on newer machines we can use TCMU.