search

gluster / gluster-kubernetes

GlusterFS Native Storage Service for Kubernetes
Apache License 2.0
875 stars 389 forks source link

Unable to create node: New Node doesn't have glusterd running / Ubuntu 18.10 / gluster on os #565

Open christoph-puppe opened 5 years ago

christoph-puppe commented 5 years ago

Hi

been trying for hrs now. Tried a few fixes I've found here, but to no avail. MAybe you have an idea what to try next.

The whole thing is on GCE. I have installed gluster on the os and want to use this cluster for heketi in a pod.

This Error Message is not very telling. Any chance we could get a more precise log entry?

Allowing file volumes on cluster. Allowing block volumes on cluster. Creating node master-1 ... Unable to create node: New Node doesn't have glusterd running Creating node worker-1 ... Unable to create node: New Node doesn't have glusterd running Creating node worker-2 ... Unable to create node: New Node doesn't have glusterd running Creating node worker-3 ... Unable to create node: New Node doesn't have glusterd running Error loading the cluster topology. Please check the failed node or device and rerun this script.

winsopc commented 5 years ago

I met the same problem, on my ubuntu 18.04.01 LTS with 4.15 kernel. you can use ./gk-deploy -gvy to get more output, but the root cause is smae for the glusterd block service not running.

nixpanic commented 5 years ago

On Fri, Mar 08, 2019 at 10:36:56AM -0800, Winson Wang wrote:

I met the same problem, on my ubuntu 18.04.01 LTS with 4.15 kernel. you can use ./gk-deploy -gvy to get more output, but the root cause is smae for the glusterd block service not running.

gluster-block can be disabled by removing or setting GLUSTER_BLOCKD_STATUS_PROBE_ENABLE to 0 in the glusterfs-daemonset. See https://github.com/gluster/gluster-kubernetes/blob/master/deploy/kube-templates/glusterfs-daemonset.yaml#L30 for more details.

In case gluster-block does not work on Ubuntu, you may want to report that as an issue at https://github.com/gluster/gluster-block

pilgylis commented 5 years ago

I got the same issue and luckily managed to get around it...

The situation is that I am deploying heketi on k8s but with existing gluster cluster external to k8s.

There are 2 issues:

  1. heketi is always use kubernetes as executor, although I did modify the value of heketi.json in a secret
  2. no private_key provided for the passwordless ssh to the gluster nodes

For the (1), turns out that there is an env variable in the deployment that overrides the value in the heketi.json, so I modified that to 'ssh' For (2), just need to encode base 64 the ssh private key and put to the secret.