After adding +X one of our services (background workers) started working, another one - for seemingly no good reason - started failing to deploy:

bin/rails aborted!

TypeError: no implicit conversion of nil into String (TypeError)

Parser.parse(source, opts)

This came from reading a configuration file relying on the RAILS_MASTER_KEY that's definetely available on that (and other services) - as nothing had changed.

Confusingly enough this only happens to our web server.

So the situation is:

- Background job service 1: Has never experienced any problem (build or deploy)

- Background job service 2: Experienced a problem on build, that got fixed after chmod +X

• Web server: Experienced a problem on build, and after chmod +X started experiencing a deploy problem as it looks like some env variables are not making it through.

This is a long shot but: Instead of adding a chmod +X, add this env var: "RAILWAY_RUN_UID=0" and tell me how it goes.

I'll give that a go it'll take me about ~18 mins with our CI so will let you know

Nope ☹ - after adding that ENV and removing the chmod I'm back to the original issue:

RUN POSTGRES__DATABASE_HOST="dummy" POSTGRES__DATABASE_REPLICA_HOST="dummy" RELEASE_SHA=REDACTED SENTRY_AUTH_TOKEN=REDACTED SECRET_KEY_BASE_DUMMY=1 SKIP_REDIS=true ./bin/rails assets:precompile

165ms

/bin/sh: 1: ./bin/rails: Permission denied

Dockerfile:65

-------------------

64 | # Precompiling assets for production without requiring secret RAILS_MASTER_KEY

65 | >>> RUN POSTGRES__DATABASE_HOST="dummy" POSTGRES__DATABASE_REPLICA_HOST="dummy" RELEASE_SHA=$RAILWAY_GIT_COMMIT_SHA \

66 | >>> SENTRY_AUTH_TOKEN=$SENTRY_AUTH_TOKEN SECRET_KEY_BASE_DUMMY=1 SKIP_REDIS=true ./bin/rails assets:precompile

67 |

-------------------

ERROR: failed to build: failed to solve: process "/bin/sh -c POSTGRES__DATABASE_HOST=\"dummy\" POSTGRES__DATABASE_REPLICA_HOST=\"dummy\" RELEASE_SHA=$RAILWAY_GIT_COMMIT_SHA SENTRY_AUTH_TOKEN=$SENTRY_AUTH_TOKEN SECRET_KEY_BASE_DUMMY=1 SKIP_REDIS=true ./bin/rails assets:precompile" did not complete successfully: exit code: 126

Any other ideas?

@medim ?

Also having this exact issue out of nowhere. Old commits work. There are no code changes that would've caused this. Already had "RAILWAY_RUN_UID=0" and still not working. Following to see what the solution ends up being for you because I'm at a loss 👀

Just to clarify on this one -https://station.railway.com/questions/bin-bash-line-1-bin-rails-permission-2994729e#5jcp - when we have the execs chmodded the deploy step that's failing is a pre-deploy. It appears as though env variables are missing in the pre-deploy but not in deploy

I'm getting almost the exact same error: "/bin/sh: 1: ../run: Permission denied". No recent changes to code that would've caused those, nor am I trying to deploy any code remotely related to the error.

I’m facing the same issue. No changes on the deploy config or relevant code changes that could impact on that.

So switching to "Metal" builds "fixed" this for us.

Thanks, jkogara. I also switched to "metal" and solved the issue.

Same for me, thanks guys.

Hello,

Thank you for reporting this. We are going to be looking into this.

This should be fixed. Let us know if you still have issues.