search

gluster / gluster-kubernetes

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

Glusterfs Pods Readiness and Liveness Probe Failed Because tcmu.runner failed. #581

Open Sheldor opened 5 years ago

Sheldor commented 5 years ago

I'm trying to install Gluster daemon hyperconverged on a completely fresh Kubernetes cluster on 3 CentOS 7.3 nodes, which are also the only nodes in the Kubernetes cluster(1 master and 2 nodes).

I just use the gk-deploy -g. But Glusterfs daemon deployment readiness and liveness probe failed.

I attached to the glusterfs container to see the journal logs. And got this: image tcmu started failed. There is no error or warn logs in /var/log/glusterfs/gluster-block/tcmu.log

I have already read this https://github.com/gluster/gluster-kubernetes/issues/572. But not works for me, because the line in my glusterfs-daemonset.yaml is already /lib/modules.

Sheldor commented 5 years ago

I found the reason. In the container, run tcmu-runner -d several times, I got this. kmod_module_new_from_lookup() failed to find module target_core_user

Run modprobe target_core_user on the node fixes this problem。