use ContainerKill instead of ContainerStop to StopTask

[dylrich]

I wanted to start a conversation about signal handling -- I noticed that the Docker driver was sending a kill signal to the container instead of the stop command and thought it would be nice if the Podman driver did this as well. I wasn't able to run all of the tests successfully locally on clone, so I'm unsure if this breaks some other expectation.

[hashicorp-cla-app]

All committers have signed the CLA.

[hashicorp-cla-app]

Thank you for your submission! We require that all contributors sign our Contributor License Agreement ("CLA") before we can accept the contribution. Read and sign the agreement

Learn more about why HashiCorp requires a CLA and what the CLA includes

_{Have you signed the CLA already but the status is still pending? Recheck it.}

[adriaandens]

Container stop sends the SIGTERM and SIGKILL signals after the timeout has been reached. Setting the timeout to 0 is the same as killing, if that's what you're after?

[dylrich]

Without this change, my tasks are not receiving the stop signals I am configuring. I want to send e.g. SIGQUIT, not the default kill signals. When I was investigating why this was, I noticed that the docker driver does not seem to be using the equivalent of podman container stop, but rather podman container kill to send the expected kill signal. I think it makes sense to make this driver match the docker driver's behavior so signals are sent to container as expected.

[tgross]

Thanks for the PR @dylrich!

This would allow the task.kill_signal to be specified, which is great. But then we need to follow-up with sending SIGKILL when the task.kill_timeout expires.

Because the Podman API can't do that for us with ContainerKill, you'll need to implement that here as well. As you've noticed in your comment, that'll be similar to what we've done for the in-tree docker driver here.

[tgross]

We've fixed it now, let's remove this comment!