nil pointer dereference in deployment monitor

[siennathesane]

Nomad version

Nomad v1.4.2

Operating system and Environment details

Issue

There is a nil pointer dereference when running nomad job run.

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x2 addr=0x8 pc=0x1042cf598]

goroutine 1 [running]:
github.com/hashicorp/nomad/command.(*DeploymentStatusCommand).ttyMonitor(0x14000a9f928, 0x1400083ec40, {0x140009fdad0, 0x24}, 0x0, 0x0)
	github.com/hashicorp/nomad/command/deployment_status.go:329 +0x1e48
github.com/hashicorp/nomad/command.(*DeploymentStatusCommand).monitor(0x1400083ee00?, 0x14000c26040?, {0x140009fdad0, 0x24}, 0x2?, 0x0?)
	github.com/hashicorp/nomad/command/deployment_status.go:185 +0x84
github.com/hashicorp/nomad/command.(*monitor).monitor(0x14000bc2690, {0x140009fda10, 0x24})
	github.com/hashicorp/nomad/command/monitor.go:302 +0xbc8
github.com/hashicorp/nomad/command.(*JobRunCommand).Run(0x1400052aa00, {0x1400004e070, 0x1, 0x1})
	github.com/hashicorp/nomad/command/job_run.go:376 +0x1074
github.com/mitchellh/cli.(*CLI).Run(0x140000e3540)
	github.com/mitchellh/cli@v1.1.4/cli.go:262 +0x4a8
main.RunCustom({0x1400004e050?, 0x3, 0x3})
	github.com/hashicorp/nomad/main.go:117 +0x350
main.Run(...)
	github.com/hashicorp/nomad/main.go:87
main.main()
	github.com/hashicorp/nomad/main.go:83 +0x50

Reproduction steps

Here is my waypoint.nomad job, it's using a GCP CSI volume on the backend.

job "waypoint" {
  region      = "americas"
  datacenters = ["us-central1"]
  type        = "service"

  group "server" {
    count = 1

    update {
      max_parallel     = 1
      canary           = 1
      min_healthy_time = "10s"
      healthy_deadline = "3m"
      auto_revert      = true
      auto_promote     = true
    }

    network {
      mode = "bridge"
      port "http" {}
      port "grpc" {}
    }

    service {
      name     = "waypoint"
      provider = "consul"
      tags = [
        "traefik.http.services.waypoint.loadBalancer.server.port=${NOMAD_PORT_http}",
        "traefik.http.routers.waypoint.rule=Host(\"waypoint.domain.io\")",
        "traefik.http.routers.waypoint.entrypoints=websecure",
        "traefik.http.routers.waypoint.service=waypoint@consulcatalog",
        "traefik.http.routers.waypoint.tls=true"
      ]
    }

    service {
      name     = "waypoint-api"
      provider = "consul"

      tags = [
        "traefik.http.services.waypoint-api.loadBalancer.server.port=${NOMAD_PORT_grpc}",
        "traefik.http.routers.waypoint-api.rule=Host(\"waypoint.domain.io\")",
        "traefik.http.routers.waypoint-api.entrypoints=websecure",
        "traefik.http.routers.waypoint-api.service=waypoint-api@consulcatalog",
        "traefik.http.routers.waypoint-api.tls=true"
      ]
    }

    volume "waypoint" {
      type            = "csi"
      source          = "waypoint"
      access_mode     = "single-node-writer"
      attachment_mode = "file-system"
    }

    task "disk-check" {
      driver = "docker"
      config {
        image   = "busybox:latest"
        command = "sh"
        args = [
          "-c",
          "chown -R 100:1000 /data/"
        ]
      }
      resources {
        cpu        = 100
        memory     = 100
        memory_max = 150
      }
      restart {
        attempts = 2
        interval = "3s"
        mode     = "fail"
      }
      lifecycle {
        hook    = "prestart"
        sidecar = false
      }
      volume_mount {
        volume      = "waypoint"
        destination = "/data"
      }
    }

    task "server" {
      driver = "docker"

      config {
        image = "hashicorp/waypoint:0.10.3"

        args = [
          "server",
          "run",
          "-accept-tos",
          "-db=/data/data.db",
          "-tls-cert-file=/home/waypoint/tls.crt",
          "-tls-key-file=/home/waypoint/tls.key",
          "-advertise-addr=https://waypoint.domain.io",
          "-listen-grpc=0.0.0.0:${NOMAD_PORT_grpc}",
          "-listen-http-insecure=0.0.0.0:${NOMAD_PORT_http}"
        ]

        ports = [
          "http",
          "grpc"
        ]

        volumes = [
          "local/tls.crt:/home/waypoint/tls.crt",
          "local/tls.key:/home/waypoint/tls.key"
        ]
      }

      template {
        data        = <<EOF
-----BEGIN CERTIFICATE-----
...
-----END CERTIFICATE-----
EOF
        destination = "local/tls.crt"
      }

      template {
        data        = <<EOF
-----BEGIN RSA PRIVATE KEY-----
...
-----END RSA PRIVATE KEY-----
EOF
        destination = "local/tls.key"
      }

      resources {
        cpu        = 100
        memory_max = 512
      }

      restart {
        attempts = 2
        mode     = "fail"
      }

      volume_mount {
        volume      = "waypoint"
        destination = "/data"
      }
    }
  }
}

And the waypoint.volume.nomad definition:

# volume registration
type      = "csi"
id        = "waypoint"
name      = "waypoint"
plugin_id = "gcepd"

mount_options {
  fs_type     = "ext4"
  mount_flags = ["noatime"]
}

capability {
  access_mode     = "single-node-writer"
  attachment_mode = "file-system"
}

topology_request {
  required {
    topology {
      segments { "topology.gke.io/zone" = "us-central1-a" }
    }
  }
}

Expected Result

The CLI wouldn't panic.

Actual Result

Nomad is throwing the same errors #13450, specifically volume max claim reached.

Job file (if appropriate)

Nomad Server logs (if appropriate)

Nomad Client logs (if appropriate)

[siennathesane]

Waypoint was failing to deploy due to a CSI issue, but the Nomad CLI was throwing this error.

[jrasell]

Hi @mxplusb and thanks for raising this issue. We will take a look into reproducing this and raising a fix.