Following the instructions here: democratic-csi/nomad.md at master · democratic-csi/democratic-csi · GitHub I was able to get some NFS volumes created.
The CSC command worked fine. I can create, list and delete volumes just fine.
I registered volumes using terraform, as suggested, and I had actions in the node storage plugins.
What didn’t work is the actual mount. Allocations took forever, and it seemed to be because there was a network timeout that prevented the nfs mount command from actually working. I’m not really sure how to figure this part out. I was using the bridged network mode for the containers, as I am not going to expose any of their ports outside of the cluster. Instead I was going to use a consul connect gateway into that app group, and then that one nginx will expose things outside the cluster.
I could get the nginx to start up, but nothing else would. The volumes would never mount. I had logs showing that it ran the NFS mount command, and it looked fine. I could run the nfs mount command by hand on the nomad client host, so it wasn’t malformed.
Is there a network isolation issue? I’m not sure what logs would be useful at the moment, I can paste my job, and I can set up another run of the whole thing to try to get more logs of the actual NFS failure.
Thanks in advance