dead jobs switch their state back to pending when you shutdown client node

[cigdono]

Nomad Version

nomad version
Nomad v0.3.2

Operating System and Environment Details

CentOS7 3.10.0-327.13.1.el7.x86_64

Issue

If I shutdown client nodes, jobs that were in the dead state switch back to the pending state.

Reproduction Steps

Configure a 3 node server cluster. Bring up a single client node. Submit a single batch job with a task count of 1 using the Docker driver. Let the job run to completion. If you run a nomad status the jobs will be in the dead state. Shutdown the client node. The job switches back to the pending state.

Note: I did not test this with other job types or drivers. May be specific to this combination... or may not be.

Nomad Server Config (other 2 are identical except for advertise IP)

bind_addr = "0.0.0.0"
log_level = "DEBUG"

data_dir = "/var/log/nomad/data_dir"
disable_update_check = true
region = "global"
datacenter = "dc1"
leave_on_terminate = true
leave_on_interrupt = true

telemetry {
    statsd_address = "127.0.0.1:8125"
}

advertise {
    http = "10.132.0.9:4646"
    rpc  = "10.132.0.9:4647"
    serf = "10.132.0.9:4648"
}

server {
    enabled = true
    bootstrap_expect = 3
    node_gc_threshold = "1m"
    retry_join = ["nomad-svr-global-01","nomad-svr-global-02","nomad-svr-global-03"]
}

Nomad Client Config

bind_addr = "0.0.0.0"
log_level = "DEBUG"
region = "global"
datacenter = "dc1"

data_dir = "/var/log/nomad/data_dir"
disable_update_check = true
leave_on_terminate = true
leave_on_interrupt = true

telemetry {
   statsd_address = "127.0.0.1:8125"
}

client {
   enabled = true
   servers = ["nomad-svr-global-01","nomad-svr-global-02","nomad-svr-global-03"]
   options = {
      "driver.raw_exec.enable" = "1"
      "docker.cleanup.image" = false
   }
}

Job File

{
    "Job": {
        "Region": "global",
        "ID": "test-01",
        "Name": "test-01",
        "Type": "batch",
        "Priority": 50,
        "Datacenters": [
            "dc1"
        ],
        "TaskGroups": [
            {
                "Name": "test-group",
                "Count": 1,
                "Tasks": [
                    {
                        "Name": "hello-world",
                        "Driver": "docker",
                        "Config": {
                            "image": "https://docker-cache.service.consul:5000/cdi/nomad-test:v0.0.9",
                            "command": "/opt/test/bin/test_batch.py",
                            "args": ["-t","30"],
                            "network_mode": "host"
                        },
                        "Resources": {
                            "CPU": 2500,
                            "MemoryMB": 256,
                            "DiskMB": 300,
                            "IOPS": 0
                        },
                        "LogConfig": {
                           "MaxFiles": 10,
                           "MaxFileSizeMB": 10
                        }
                    }
                ]
            }
        ]
    }
}

[dverbeek84]

I see a sort of problem with the Docker driver.

When the Docker container is dead and you restart Nomad. Nomad wil not remove the dead container with nomad stop <job>

[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.