Make start.sh abort and fail if one of its steps failed

[vorburger]

This seems like a good idea, and could reduce confusion?

One example, among many different possible failures obviously, is what I ran into in #115: nginx failed to start, but the shim ran anyway. That's confusing, and it's probably safer to "fail fast" and abort on an emerg from Nginx, or any similar launch failures.

Please note that I HAVE NOT TESTED THIS 😈 at all... Does this project run integration tests on PRs?

Would you need me to rebuild and run with this to make sure it's fine? Or can you test it?

[joaosa]

This seems like a good idea, and could reduce confusion?

One example, among many different possible failures obviously, is what I ran into in #115: nginx failed to start, but the shim ran anyway. That's confusing, and it's probably safer to "fail fast" and abort on an emerg from Nginx, or any similar launch failures.

Please note that I HAVE NOT TESTED THIS 😈 at all... Does this project run integration tests on PRs?

Would you need me to rebuild and run with this to make sure it's fine? Or can you test it?

Thanks @vorburger! Your approach makes sense and I would even extend it to set -euo pipefail 👍

I'll try to test this when I get a chance. This project does run integration tests, but they do not cover all aspects of the codebase :)

[vorburger]

I would even extend it to set -euo pipefail +1

done!