Hi,

Please add next build to your build command.

buildCommand = "cd packages/convex-backend && npx convex deploy --cmd 'cd ../../ && pnpm build:starter-convex' --cmd-url-env-var-name NEXT_PUBLIC_CONVEX_URL"

i do have it --> pnpm build:starter-convex in the above and then

startCommand = "pnpm start:starter-convex"

buildCommand = "cd packages/convex-backend && npx convex deploy --cmd 'cd ../../ && pnpm build:starter-convex' --cmd-url-env-var-name NEXT_PUBLIC_CONVEX_URL"

i do have it --> pnpm build:starter-convex in the above and then

startCommand = "pnpm start:starter-convex"

What does build:starter-convex do?

I have 1 more doubt, Is it safe to upload logs? I can help upload the logs success one with nixpacks and another with RAILPACK which failed.

I hope it does not have any sensitive information. Could you please confirm.

Yes, it is safe to upload logs.

"build:starter-convex": "turbo build --filter=starter-convex",

Since its a turbo repo it will build the next app with shared packages as well.

Do you see just a build part of your script anywhere? It keeps chaining down to that command.

https://github.com/varunteja007006/open_templates/tree/main/my-turbo-app/logs

The logs

Its actually confusing because when I set builder = "nixpacks" in .toml it works but not when I set it to builder = "RAILPACK".

I will try to debug that build sequence via logs and if possible ssh

I would try to deploy it myself but Railway has horrible caching policies and I won't be able to deploy this repository for a while.

Can you try setting your start command to be sleep 999999, ssh into the container, and try to find where the .next directory or any build directory is located?

I was able to ssh into the successful part of deploy not the failed one. Is it possible to ssh into the failed container?

sleep 999999 --> in the .toml file?

https://gemini.google.com/share/fa8d739c601d

Made an analysis with AI.

Thanks that sleep command did help i dont see the .next folder when I ssh into the container. Its strange I am not sure if I have to update my deploy logic or something wrong with the way RAILPACK works. What do you suggest?

Had to update the script for it to work. Must be the way railpack reads the script i guess.

Thanks a lot Sam

Glad to hear you got it fixed!