Hey! I am having a weird issue with the Docker environment on OSX. Some times my tests take a really long time, and in the end I am just served with this weird error message:

15:23:31.86 [INFO] Completed: Building pytest_runner.pex
15:23:31.87 [ERROR] 1 Exception encountered:

Engine traceback:
  in `test` goal

IntrinsicError: Failed to start Docker container `510f97a829fa18c76691db4e556d07e31e16c55eea3aee870305a2779346ef6b` for image `sha256:e4bf78b64f77acd1f2201a360900c645575cd5aefdf46d4cc5025194651f30ac`: RequestTimeoutError

The docker container that pantsd uses (?) to avoid having to reinitialize the environment every time seems unresponsive. I can't kill it with for example docker kill:

(requirements.pex) peder(feature/azure-devops-api)$ docker kill cranky_davinci
Error response from daemon: Cannot kill container: cranky_davinci: tried to kill container, but did not receive an exit event
(requirements.pex) peder(feature/azure-devops-api)$

Nothing works at this point, so I try to restart docker, but then Pants complains (rightfully so), that its docker container is missing:

Failed to create Docker execution in container: DockerResponseServerError { status_code: 409, message: "Container 78d6a27f5e7c47ac676b0a42923670b049cc86d82c274a8e47b2a1006ddf9844 is not running" }

My questions are: • Why would the container become stuck in this weird state? This might be due to my environment, I sometimes have the wrong docker context set which causes docker to fail out for Pants some times • How would I let pantsd know to recreate the container instead of just complaining about it being missing?

Hi, sorry for the trouble. Are you able to create a small example repo that reproduces this problem?

The second point is https://github.com/pantsbuild/pants/issues/18209 . The workaround I found is to kill pantsd only:

pkill -f pantsd

(This doesn’t help with the first/fundamental issue, though)

Hey Benjy! Yeah sure, I'll try my best, thank you for reaching out.