Hello, I am seeing failing deploys (the fly command exits non-zero) with the “Error 422 Unprocessable Entity” message at the end. Here’s a snippet of the last bit of output for this app (fe-anchor-stg-anchordotdev):
--> Pushing image done
==> Creating release
--> release v129 created
--> You can detach the terminal anytime without stopping the deployment
==> Release command detected: bin/rails db:reset db:migrate db:seed
--> This release will not be available until the release command succeeds.
Error 422 Unprocessable Entity
I encountering this both locally and in a github action.
release command fail, stop the progress.
The release command is weird, db:reset will drop the db and recreate it, it should not use in production env. Usually db:migrate is enough.
What is your deployment method? images or buildpacks?
It’s a staging app that reseeds the database on every deploy. On a normal failure, i get something like the following:
Tasks: TOP => db:migrate:with_data
(See full trace by running task with --trace)
Main child exited normally with code: 1
Main child exited normally with code: 1
Starting clean up.
Starting clean up.
Error release command failed, deployment aborted
I don’t see any exit code or indication of an error, and the deploy seems to have succeeded despite the “Error 422” message. Is this a problem with the platform or something in my own deploy scripts?
/cc @jsierles @eli
Migrate task fails, I’m not sure how to check release logs, maybe try flyctl logs.
Can you try again with LOG_LEVEL=debug and post the result? This will help determine which request is getting the 422 response.
Thanks for the LOG_LEVEL tip, I’ve added it to my deploy script but i’m not seeing the 422 response anymore. I’ll reply here with the debug output if it happens again.