Modifying device count doesn't cause scheduling

[cgbaker]

Nomad version

Nomad 0.9.0, 0.9.6, 0.10.1

Operating system and Environment details

MacOS, example device plugin

Issue

After updating the device count in the resources stanza, the scheduler does not detect the need for a destructive allocation update, and the updated device count is not reflected in the updated allocation.

The reason is, at least, that the task update check does not consider the new device field:
https://github.com/hashicorp/nomad/blob/v0.10.1/scheduler/util.go#L399-L404

Reproduction steps

1. Run job and note allocation resources:

Device Stats
hashicorp/example/hello[hello-1] 1

2. Modify job to increase device "example" count, note that allocation is updated:

Allocation "0abefd49" modified: node "a1a38f1b", group "cache"

3. Examine allocation resources, no new device was attached.

Job file (if appropriate)

job "example" {
  datacenters = ["dc1"]
  group "cache" {
    task "redis" {
      driver = "docker"
      config {
        image = "redis:3.2"
      }
      resources {
        device "example" {
          count = 1
        }
      }
    }
  }
}

Nomad Server logs (if appropriate)

On job update:

    2019-11-07T16:47:30.580Z [DEBUG] worker.service_sched: reconciled current state with desired state: eval_id=aba15a40-2bb7-74f7-8b47-651f81445050 job_id=example namespace=default results="Total changes: (place 0) (destructive 0) (inplace 1) (stop 0)
Desired Changes for "cache": (place 0) (inplace 1) (destructive 0) (stop 0) (migrate 0) (ignore 0) (canary 0)"

[github-actions]

I'm going to lock this issue because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.