job resource not returning deployment_id or deployment_status

[cgbaker]

Terraform Version

Terraform v0.12.25
Nomad Provider Version 1.4.5

Nomad Version

0.11.2

Issue

When setting a job resource with detach = false, the deployment information (if any) should be returned in deployment_id and deployment_status. This is not happening because of an improper return from deploymentStateRefreshFunc.

Furthermore, the provider was not unsetting these when detach was changed to true.

[mahsoud]

@cgbaker What is the expected behaviour now for batch jobs? I started getting 404 error with the release of this fix:

resource nomad_job this {
jobspec = file("${path.module}/nomad_job.tpl")
deregister_on_destroy = true
deregister_on_id_change = true
detach = false
}

$ terraform apply
nomad_job.this: Creating...
nomad_job.this: Still creating... [10s elapsed]
nomad_job.this: Still creating... [20s elapsed]
nomad_job.this: Still creating... [30s elapsed]
Error: error waiting for job 'my-job-name' to deploy successfully: Unexpected response code: 404 (eval not found)

[cgbaker]

That's a bug on my part. I'll fix that today. My apologies.

…

On Tue, May 19, 2020, 09:02 mahsoud ***@***.***> wrote: @cgbaker <https://github.com/cgbaker> What is the expected behaviour now for batch jobs? I started getting 404 error with the release of this fix: resource nomad_job this { jobspec = file("${path.module}/nomad_job.tpl") deregister_on_destroy = true deregister_on_id_change = true detach = false } $ terraform apply nomad_job.this: Creating... nomad_job.this: Still creating... [10s elapsed] nomad_job.this: Still creating... [20s elapsed] nomad_job.this: Still creating... [30s elapsed] Error: error waiting for job 'my-job-name' to deploy successfully: Unexpected response code: 404 (eval not found) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#104 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAMY6T3BKMHY3Z27SER54X3RSKGOVANCNFSM4NCRXT5Q> .

[cgbaker]

@mahsoud, would you mind filing a new issue? I'd like to see the Terraform output with TF_LOG=8, as well as the Nomad logs.

It seems that in your case, the evaluation is being garbage collected while waiting for the deployment (it shouldn't be waiting, because batch jobs don't create deployments.) So, the provider needs to better handle the GC'd evaluation, and it needs to handle the bug around the deployment wait (which was apparently the result of a change in the Plugin SDK).

[mahsoud]

the job gets executed very quickly, and yes in some cases the allocation is GC very quickly.

regardless, thank you! We'll file the issue shortly