Fork process worker found in dead state when attempting to run login lookup plugin

[artis3n]

When trying to use community.hashi_vault.vault_login, I am receiving the following error:

TASK [Vault | Login] ******************************************************************************************************
objc[35867]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called.
objc[35867]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
ERROR! A worker was found in a dead state

This error is consistent between community.hashi_vault.vault_login and community.hashi_vault.vault_token_create.

ansible-galaxy collection list tells me I am using community.hashi_vault 5.0.0. I also have hvac 1.1.0 in my poetry's pyproject.toml. I have tried reverting to hvac 0.11.2 but still get the error.

My task is:

- name: Vault | Login
  ansible.builtin.set_fact:
    hashivault_login_data: "{{ lookup('community.hashi_vault.vault_login', url='http://vault:8200', auth_method='approle', role_id='{{ vault_role_id }}', secret_id='{{ vault_secret_id }}') }}"

Where http://vault:8200 is a server in an intranet. I've confirmed I can vault write auth/approle/login ... with the role ID and secret ID I'm attempting to use and login against this vault server works fine with the CLI call.

The role ID and secret ID are injected with -e during the ansible run, and I've verified they contain the correct contents with ansible.builtin.debug calls.

ansible-playbook -i inventory -e vault_role_id=... -e vault_secret_id=...

I have no idea why I'm getting this error. The host I'm running the playbook on is OSX M1.

[briantist]

Thank you for including the OS; my first question was going to be if you're on a mac since forking errors with ansible tend to be common there. I don't think this error is actually related to the collection.

You can try setting --forks=1 just to see if the issue still persists or not.

Here's two old issues that talk about it a bit (though one of them it more pertinent to connecting to Windows hosts, and your issue is in a set_fact so not even a module):

• ERROR! A worker was found in a dead state ansible/ansible#32554
• ERROR! A worker was found in a dead state ansible/ansible#32554

If the issue persists, I would recommend raising it in real-time chat over Matrix or IRC:

• https://docs.ansible.com/ansible/latest/community/communication.html#real-time-chat

The general usage and support channel is probably the right place:

• https://matrix.to/#/#users:ansible.com

I'll leave this open for now, in case there's anything that can be done but based on previous experience I don't think that's possible.

Please post an update if you find anything new, thanks!

[artis3n]

Thanks for the quick and comprehensive response! The forks parameter didn't help (for anyone who comes to this in the future), but thanks to those resources I was led to https://www.whatan00b.com/posts/debugging-a-segfault-from-ansible/ and using no_proxy='*' fixes the apparently underlying Python and Ansible and OSX issue.

I'm now getting an auth error that doesn't make sense because I can vault write a successful auth with the creds, but that is well within my own troubleshooting and doesn't appear that anything is wrong in the plugin.

fatal: [snuc]: FAILED! => {"msg": "An unhandled exception occurred while running the lookup plugin 'community.hashi_vault.vault_login'. Error was a <class 'hvac.exceptions.InvalidRequest'>, original message: invalid role ID, on post http://vault:8200/v1/auth/approle/login. invalid role ID, on post http://vault:8200/v1/auth/approle/login"}