Burak Yigit Kaya

@DuncanConroy `podman compose run` calls add a weird hash at the beginning (see https://github.com/getsentry/self-hosted/actions/runs/14732417876/job/41349793317?pr=3673#step:4:890) Any idea how to suppress that?

FYI I put this on hold on my end folks as I spent way too much time fixing these scripts and CI. If anyone can figure out why we are...

@doc-sheet you mean GitHub CI debug? Because I'm fairly sure the debug for the tooling will generate output that would break things even further. Btw the setup is fairly bare...

Could that issue be related to https://github.com/containers/podman-compose/issues/707 ?

@doc-sheet here's the run with debug enabled: https://github.com/getsentry/self-hosted/actions/runs/14732867860/job/41748307991?pr=3673 Thanks so much for your help 🙏🏻

@doc-sheet ah, great find! The `$dcr` shortcut is only for one-off commands so disabling pod creation for that makes the most sense to me. I'll make the change and see...

We should also figure out volume caching story with Podman. We have 2 options: - Use `podman volume export` and `podman volume import`. This looks neat but requires us to...

Okay, now we need to make the tests aware of Podman

@aldy505 the tests written in Python assume everything with Docker and Docker Compose. They now need to be aware of what technology we're using. See the fail here: https://github.com/getsentry/self-hosted/actions/runs/14893466765/job/41831036774?pr=3673#step:4:2280

So apparently this is a thread-safety issue: https://github.com/ros-visualization/rqt_robot_monitor/issues/6 @Kriechi we can consider adding a lock into the `search` method or make it work over a copy. Looking at the stack...