Cannot get good status curl result after build

[cobinrox]

Greetings- I've been playing/hacking for several hours to get just up through the curl command in the getting started doc. I'm running with GitBash on Windows and can run make build and the make upd commands ok.

But when i run the curl to get overall status I get this:
$ curl -s http://localhost:8081/healthz <br /> <b>Warning</b>: require_once(/app/vendor/autoload_runtime.php): Failed to open stream: No such file or directory in <b>/app/public/index.php</b> on line <b>5</b><br /> <br /> <b>Fatal error</b>: Uncaught Error: Failed opening required '/app/vendor/autoload_runtime.php' (include_path='.:/usr/local/lib/php') in /app/public/index.php:5 Stack trace: #0 {main} thrown in <b>/app/public/index.php</b> on line <b>5</b><br />

I was able to get a docker exec to install some necessary packages, and finally got the status to come back ok afterward, but I wanted to get a clean, consistent build and run so started hacking at the Makefile so that anyone trying a clean install could get the right installations out-of-the-box. But, that became a rabbit hole since I don't know the entire architecture of the various make/compose files. So, before going down that rabbit hole, I just wondered if this curl error was indeed expected behavior for a clean install?

Here's the after-install ad-hoc command to get extras installed, btw:
docker compose exec backend-symfony composer install

[sueun-dev]

That fatal is expected on a fresh clone, nothing you broke. backend-dev bind-mounts ./backend-symfony:/app, which shadows the vendor/ the Dockerfile builds into the image, and vendor/ is gitignored so the host clone has none. So index.php:5 can't load autoload_runtime.php until you populate the mounted dir.

Your composer install is the right fix (or make composer-install); on 858bb74, healthz went back to {"status":"ok"} right after for me. The getting-started doc just runs the healthz check one step before composer-install.

[laugiov]

Thanks for the report, and sorry for the hours you lost on it. This was a real gap in our docs, not something you broke.
suen-dev has the diagnosis exactly right. The backend-dev service mounts ./backend-symfony over /app so local edits are live without a rebuild, and that mount hides the vendor/ directory built into the image. Since vendor/ is gitignored, a fresh clone has nothing behind the mount, and public/index.php fails on line 5. The getting started doc placed the healthz check one step before the install, so you hit it at the worst possible moment.
Fixed on main (#64):
make up and make upd now install the dependencies on the first run, only when vendor/autoload_runtime.php is missing. No cost after that.
make doctor now checks for it and tells you to run make composer-install.
The healthz check moved after the install step in the doc, and the error message is in the troubleshooting section.