issue in testing with debian 11 aka bullseye

[zerwes]

@zerwes If there is an issue in testing with debian11, would you mind filing an issue for it, with the logs? We'll try to figure it out. It's probably a molecule thing again.

Originally posted by @evrardjp in #189 (comment)

[zerwes]

while running molecule test on debian 11 with

$ molecule --version
molecule 3.5.2 using python 3.9 
    ansible:2.12.1
    delegated:3.5.2 from molecule
    docker:1.1.0 from molecule_docker requiring collections: community.docker>=1.9.1

I get in the converge step
INFO Running default > converge
a error related to systemd not beeing available in the docker image

TASK [ansible-keepalived : Check if keepalived is enabled] *********************
fatal: [keepalived-bionic]: FAILED! => {"changed": false, "cmd": "systemctl daemon-reload && systemctl is-enabled keepalived.service", "delta": "0:00:00.004321", "end": "2022-01-20 19:34:58.212965", "failed_when_result": true, "msg": "non-zero return code", "rc": 1, "start": "2022-01-20 19:34:58.208644", "stderr": "Failed to connect to bus: No such file or directory", "stderr_lines": ["Failed to connect to bus: No such file or directory"], "stdout": "", "stdout_lines": []}

(seems to fail on every image i.e.:

TASK [ansible-keepalived : Check if keepalived is enabled] *********************
fatal: [keepalived-centos7]: FAILED! => {"changed": false, "cmd": "systemctl daemon-reload && systemctl is-enabled keepalived", "delta": "0:00:00.469191", "end": "2022-01-20 20:55:46.937047", "failed_when_result": true, "msg": "non-zero return code", "rc": 1, "start": "2022-01-20 20:55:46.467856", "stderr": "Failed to get D-Bus connection: No such file or directory", "stderr_lines": ["Failed to get D-Bus connection: No such file or directory"], "stdout": "", "stdout_lines": []}

I thing this is related to the geerlingguy images ... (I can recall that in another context I had some issues with them too), but docker and molecule are not my area of expertise, so this might be only half of the truth ...

On the branch fix-docker I have created a working molecule scenario with a self build docker image with debian 11 and systemd working on debian 11 ...

I try to find out some more details if I find the time ...

[evrardjp]

Thanks for the first pass, and first documentation!
I am not running debian myself, so it makes it a bit harder.

I however will highlight that you're looking for help.

[evrardjp]

I will try to find time to compare, and maybe fix a few things.

For the images, it's been a back and forth between different images in the past.
I think it would make sense to fix the upstream images rather than carrying our own.

[zerwes]

I think it would make sense to fix the upstream images rather than carrying our own.

I fully agree to this, the commit was just a example. As I mentioned, is stumbled over the problems with the geerlingguy images before, but did not find the time to dive deeper into it.
As those images are quite popular, maybe the error is in my docker setup (as I mentioned, I am rather a simple user on that topic)
If I can provide some more information, just let me know.

[zerwes]

opened geerlingguy/docker-debian11-ansible#4

[evrardjp]

TBH, I had issues also in the past with those images. I brought my own images, but it was a burden to maintain.

[zerwes]

I fully agree ... for a q&d test some self build images are totally OK. With the full implication the burden you mentioned ...
But as the geerlingguy images are really widely used, I think the intention to fix them upstream (or at least to understand why they fail here) is more in the interest of the open source idea then everybody brewing it's own beer.
So if you have any idea, what I can test with a reasonable effort and expense of resources just let me know.

[evrardjp]

I fully agree with you.

If I have an idea, I will let you know.

Let's sync over the issue you opened upstream. In the meantime, you can use github actions to test your patches.

[evrardjp]

I recently updated our OS list in #215 , and I am using the images from Jeff Geerling again.
It seems the problem was solved again...

I think this is fine now because:

1. We have more test coverage by adding debian into the default tests
2. We'll run more regular jobs to test if everythins is still fine -> An issue could be found earlier.

Let's close this and reopen should the issue arise.