search

gluster / gluster-kubernetes

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

Not working against a fresh OpenShift cluster #258

Open paralin opened 7 years ago

paralin commented 7 years ago

I'm running OpenShift Origin on four bare metal CentOS 7 servers. I've been trying to get this to work for days now.

Always this happens:

Using OpenShift CLI.
NAME        STATUS    AGE
glusterfs   Active    3m
Using namespace "glusterfs".
Waiting for deploy-heketi pod to start ... OK
Found node bart on cluster 7eaa67c5087f92b7bd64b349d44ead8c
Found device /dev/loop0
Creating node lisa ... Unable to create node: Unable to execute command on glusterfs-6ddhz:
Creating node marge ... Unable to create node: Unable to execute command on glusterfs-6ddhz:
Creating node homer ... Unable to create node: Unable to execute command on glusterfs-6ddhz:
Error loading the cluster topology.
Please check the failed node or device and rerun this script using the --load option.

I verified that the names of the "manage" nodes are exactly what they are in my oc nodes output. I've added the privileged SCC to the default and router service accounts in the namespace. At this point I'm not sure what else to try.

paralin commented 7 years ago

I traced the issue to:

Logs here:

glusterd[51]: pending frames:                                                                                                                               
systemd[1]: glusterd.service: main process exited, code=killed, status=11/SEGV
MohamedAshiqrh commented 7 years ago

@paralin can you specify the gluster image name and tag?