ansible output

[hwts]

The output of ansible is not displayed.

instead

Plan: 1 to add, 0 to modify, 0 to destroy.
ansible_playbook.zabbix-server: Creating...
ansible_playbook.zabbix-server: Still creating... [10 s elapsed].
ansible_playbook.zabbix-server: Still creating... [20s elapsed].
ansible_playbook.zabbix-server: Still being created... [30s elapsed]
ansible_playbook.zabbix-server: Still creating... [40s elapsed]
ansible_playbook.zabbix-server: Creation completed in 43s [id=2023-07-19 01:47:33.374074 +0300 m=+0.120067126]

[lae]

Providers don't generally send output to Terraform during provisioning.

You can capture the playbook output and expose it via the ansible_playbook_stdout output from the resource: https://registry.terraform.io/providers/ansible/ansible/latest/docs/resources/playbook#ansible_playbook_stdout

[smtarslanturk]

Please try codes like below Change variables with your resource name

output "playbook_stderr" {
  value = ansible_playbook.<RESOURCE_NAME>.ansible_playbook_stderr
}

output "playbook_stdout" {
    value = ansible_playbook.<RESOURCE_NAME>.ansible_playbook_stdout
}

@hwts

[hwts]

This is the design that allows you to see the process of making a playbook

  provisioner "local-exec" {
      working_dir = "./ansible/"
      command = "ansible-playbook -u ${var.ssh_user} --key-file ${var.ssh_keys["priv"]} -i inventory site.yml"
  }  

[baznikin]

It can work, if resources was created. In my case resource creation failed because of some error and there is absolutely no way to see error, only to run ansible-playbook manually. TF_LOG=TRACE do not show provider output too

[Vasil-Derilov]

I have an issue with that as well. Lets say the playbook gets stuck on something and won't time-out. I have no idea of what caused it. This makes playbooks really hard to debug.

The provisioner approach is okay but it's sub-optimal.

[minsis]

If you set ignore_playbook_failure to true, then terraform will continue to execute as if the resource was successful. This will then populate ansible_playbook_stderr and ansible_playbook_stdout.

Then in my outputs:

output "ansible_playbook_stderr" {
  value = ansible_playbook.nginx[*].ansible_playbook_stderr
}

output "ansible_playbook_stdout" {
  value = ansible_playbook.nginx[*].ansible_playbook_stdout
}

I agree it's a bit annoying as maybe this could be output with TRACE logging level. If you have more resources to execute after the playbook is supposed to run, then setting ingore_playbook_failure to true is going to break the rest of the terraform run.

[lae]

I would recommend commenting on #39 with complaints/suggestions for how to deal with playbook errors.

[gravesm]

We have improved logging with v1.2.0 of this provider. As stated above, using ignore_playbook_failure can be helpful for debugging playbook problems.