Nomad: Constraint "missing drivers" issue

[devfacet]

Hi All,

I have a dev machine that has Nomad server and client are running on it. Somehow I'm getting missing drivers issue. Any thoughts?

# Linux CentOS 6 - Kernel 3.10.93-1.el6.elrepo.x86_64

$ docker --version
Docker version 1.8.3, build f4bf5c7

$ ./nomad run example.nomad

==> Monitoring evaluation "add8b71b-89c9-b852-cffc-c0375d1d0d6f"
    Evaluation triggered by job "example"
    Scheduling error for group "cache" (failed to find a node for placement)
    Allocation "c81b4c13-6394-3078-db3d-01fa5b31b73c" status "failed" (1/1 nodes filtered)
      * Constraint "missing drivers" filtered 1 nodes
    Evaluation status changed: "pending" -> "complete"
==> Evaluation "add8b71b-89c9-b852-cffc-c0375d1d0d6f" finished with status "complete"

$ ./nomad node-status
ID                                    DC   Name                                     Class   Drain  Status
c7a2f801-da24-0795-45e8-7117a3aa0a0d  dc1  #  <none>  false  ready

[devfacet]

I was able to fix it by restarting docker daemon and nomad agent ( client )

[mongrelion]

I ran into this issue today. Happened that I didn't have Docker installed on my machine.
Perhaps the wording of the error message could be improved?

[dadgar]

@mongrelion Do you have any suggestions?

[mongrelion]

@dadgar I'm no expert in UX but I think that something like missing drivers (is blah service running?) should suffice.

[dadgar]

Yeah, docker however is the only driver that requires a service to be running.

[mongrelion]

Ah, bummer. Perhaps, then, a special note on the Getting Started guides or somewhere else, specifying that the host were the task is going to be ran must have Docker installed. I know this sounds kind of obvious but I have seen these sort of cases before with other schedulers that integrate with Docker and such.

[digikin]

I found that the issue is installing docker after nomad. So if you already have nomad installed, stopping everything ctl + c then as root restart docker systemctl restart docker. Once docker reboots turn on nomad again and the error disappears.