Submit job HCL with API and show it as default in UI

Nomad
1

Hi all,

I use the /v1/jobs endpoint to submit my job, and for that I need to convert my job.hcl to job.json first and then submit, But I would like to use the job.hcl directly and show it in the UI as default as it is much compact and easier to work with. Is there a way this can be achieved? I searched web and this forum and could only find /v1/jobs/parse endpoint but, will this preserve the HCL?

Thanks.

2

hi, from Jobs - HTTP API | Nomad | HashiCorp Developer :

Submission (JobSubmission: <optional>) - Specifies the original HCL/HCL2/JSON definition of the job. This data is useful for reference only, it is not considered for the actual scheduling of Job.

I do not really see “JobSubmission” type documented. See nomad/api/jobs.go at b7419bc940151684d19054ee3cef6b732196b50a · hashicorp/nomad · GitHub .

3

Hi @ayushk , thanks for raising the question, and thanks @Kamilcuk for responding. Apologies for not having this better-documented. JobSubmission data should include at a minimum a Format string ("hcl2") and the Source as a stringified representation of your jobspec hcl. Here is an example using our hello-world job template:

{
  "Job": {
    "Affinities": null,
    "AllAtOnce": false,
    ...
    "VaultToken": "",
    "Version": 0,
    "meta": {}
  },
  "Submission": {
    "Source": "job \"hello-world\" {\n  # Specifies the datacenter where this job should be run\n  # This can be omitted and it will default to [\"*\"]\n  datacenters = [\"*\"]\n\n  meta {\n    # User-defined key/value pairs that can be used in your jobs.\n    # You can also use this meta block within Group and Task levels.\n    foo = \"bar\"\n  }\n\n  # A group defines a series of tasks that should be co-located\n  # on the same client (host). All tasks within a group will be\n  # placed on the same host.\n  group \"servers\" {\n\n    # Specifies the number of instances of this group that should be running.\n    # Use this to scale or parallelize your job.\n    # This can be omitted and it will default to 1.\n    count = 1\n\n    network {\n      port \"www\" {\n        to = 8001\n      }\n    }\n\n    service {\n      provider = \"nomad\"\n      port     = \"www\"\n    }\n\n    # Tasks are individual units of work that are run by Nomad.\n    task \"web\" {\n      # This particular task starts a simple web server within a Docker container\n      driver = \"docker\"\n\n      config {\n        image   = \"busybox:1\"\n        command = \"httpd\"\n        args    = [\"-v\", \"-f\", \"-p\", \"${NOMAD_PORT_www}\", \"-h\", \"/local\"]\n        ports   = [\"www\"]\n      }\n\n      template {\n        data        = <<-EOF\n                      <h1>Hello, Nomad!</h1>\n                      <ul>\n                        <li>Task: {{env \"NOMAD_TASK_NAME\"}}</li>\n                        <li>Group: {{env \"NOMAD_GROUP_NAME\"}}</li>\n                        <li>Job: {{env \"NOMAD_JOB_NAME\"}}</li>\n                        <li>Metadata value for foo: {{env \"NOMAD_META_foo\"}}</li>\n                        <li>Currently running on port: {{env \"NOMAD_PORT_www\"}}</li>\n                      </ul>\n                      EOF\n        destination = \"local/index.html\"\n      }\n\n      # Specify the maximum resources required to run the task\n      resources {\n        cpu    = 50\n        memory = 64\n      }\n    }\n  }\n}",
    "Format": "hcl2"
  }
}

See the struct that Kamilcuk linked above for optional variables as well.