Man, I can't figure out why tracy isn't correctly producing a log. Sometimes it works, sometimes it doesn't. Really it doesn't seem set up for automated deployments like this. I think right now the issue is it's ignoring a SIGINT, which is supposed to shut it down and write out the log, but possibly there's also another issue related to spacetime server restarts. I'll try rearranging the script to let it shut down on its own...

It looks like the issue is that spacetime is failing to reboot. Currently, the script:

1. runs spacetimedb without tracy to upload the world
2. uploads the world
3. shuts down spacetimedb
4. starts spacetimedb with tracy
5. runs tracy
6. runs a bot test
7. shuts down spacetimedb and tracy
8. zips up all the logs and tracy trace and uploads them

the problem right now seems to be step 4. Sometimes spacetimedb is failing to restart after being shutdown via a SIGINT (CTRL-C in bash.) The error is the following:

Error: database error: IO error: could not acquire lock on "/home/ubuntu/actions-runner-linux-x64-2.314.1/_work/SpacetimeDB/SpacetimeDB/crates/bench/bottest/.spacetime/control_node/control_db/db": Os { code: 11, kind: WouldBlock, message: "Resource temporarily unavailable" }

Not totally sure what to do next here. I could try and avoid restarting spacetimedb and just run tracy the whole time the world is uploading, but then the tracy logs are going to get huge.

Closing in favor of #1255