Skip to content

EAS Hosting deploy --prod fails with "Unexpected server error" when promoting deployment to production #4648

Open

Description

@Lokoloko1444

Build/Submit details page URL

No response

Summary

Hi Expo Support,

I'm experiencing a persistent issue with EAS Hosting on my project and I've exhausted all documented workarounds. I'd like to request investigation and, if needed, help migrating my project to the correct account.

Project details

  • Project: app-bilddin
  • Project ID: 23eda9b4-13b8-4659-8896-d5bbba60e477
  • Account (owner of project): lg0144
  • Currently logged in as: lgoatrt (Owner of lg0144 and lokoloko-2)
  • Custom domain: taze.to

What I'm trying to do

Deploy my web app to production so the custom domain taze.to can go live via EAS Hosting.

What happens

Every attempt to run eas deploy --prod fails at the promotion step:

Created deployment
100% Uploading 78 assets
× Failed to promote deployment to production
Unexpected server error (Expo error ID: ).
Request ID:
Error: GraphQL request failed.

I've tried multiple times over several hours, with different Expo error IDs (e.g. c2d3b7f4-a90f-4410-bf1d-493843cabd63, 7a0eb773-d2a4-4496-a1bd-351a89b06563, 865152a1-5e8d-4634-93c3-3947a944ef49, 5bc169e9-aa72-4624-91a0-f5a83d12f1a4).

What I've already tried

  1. Corrected app.json: owner → lg0144, slug → app-bilddin, projectId → 23eda9b4-13b8-4659-8896-d5bbba60e477
  2. Logged in with both lokoloko-2 and lgoatrt accounts (both Owners)
  3. Manual promotion via eas deploy:alias --prod --id=<deployment-id> — fails with "Worker Deployment with deployment identifier does not exist"
  4. Node.js 22.23.3 via fnm (upgraded from Node 20 because EAS CLI now requires Node 22+)
  5. Verified DNS records:
    • TXT _cf-custom-hostname.taze.to → 1266fffd-f4da-4b47-b167-4053ffaf5de6
    • CNAME _acme-challenge.taze.to → taze.to.7dede208f6a1d738.dcv.cloudflare.com
    • A taze.to → 172.66.0.241 (currently DNS only for validation)
  6. Removed Cloudflare Worker custom domains on taze.to and www.taze.to so EAS can take over
  7. Checked status.expo.dev — all systems operational

What I suspect

This looks like the known bug in EAS CLI where assignAlias doesn't verify the alias actually points to the new deployment, and the server mutation fails without a clear client-side error. Reference: PR #4443 in expo/eas-cli.

What I'm also requesting

Please clean up my account structure. I want:

  • Keep: lgo-studio (the paid account that should be billing and owning everything)
  • Keep: lg0144 (project owner)
  • Remove/unlink: lokoloko-2 (should no longer be involved with billing or this project)

Specifically, I want it confirmed that:

  • Billing is tied only to lgo-studio
  • The app-bilddin project is owned only by lg0144
  • lokoloko-2 has no ownership, billing, or access rights to this project

What I need from you

  1. Investigate why deploy --prod fails with a server error and let me know if there is a project-side issue I can fix
  2. If it's a server-side bug, please fix or escalate
  3. Confirm the account cleanup above (billing only on lgo-studio, lokoloko-2 removed)
  4. Once resolved, help me get taze.to live on the production deployment

Thank you.

Davy van Praet
lgstudio144@gmail.com

Managed or bare?

yes

Environment

EAS Hosting deploy --prod fails with "Unexpected server error" when promoting deployment to production

Error output

No response

Reproducible demo or steps to reproduce from a blank project

Hi Expo Support,

I'm experiencing a persistent issue with EAS Hosting on my project and I've exhausted all documented workarounds. I'd like to request investigation and, if needed, help migrating my project to the correct account.

Project details

  • Project: app-bilddin
  • Project ID: 23eda9b4-13b8-4659-8896-d5bbba60e477
  • Account (owner of project): lg0144
  • Currently logged in as: lgoatrt (Owner of lg0144 and lokoloko-2)
  • Custom domain: taze.to

What I'm trying to do

Deploy my web app to production so the custom domain taze.to can go live via EAS Hosting.

What happens

Every attempt to run eas deploy --prod fails at the promotion step:

Created deployment
100% Uploading 78 assets
× Failed to promote deployment to production
Unexpected server error (Expo error ID: ).
Request ID:
Error: GraphQL request failed.

I've tried multiple times over several hours, with different Expo error IDs (e.g. c2d3b7f4-a90f-4410-bf1d-493843cabd63, 7a0eb773-d2a4-4496-a1bd-351a89b06563, 865152a1-5e8d-4634-93c3-3947a944ef49, 5bc169e9-aa72-4624-91a0-f5a83d12f1a4).

What I've already tried

  1. Corrected app.json: owner → lg0144, slug → app-bilddin, projectId → 23eda9b4-13b8-4659-8896-d5bbba60e477
  2. Logged in with both lokoloko-2 and lgoatrt accounts (both Owners)
  3. Manual promotion via eas deploy:alias --prod --id=<deployment-id> — fails with "Worker Deployment with deployment identifier does not exist"
  4. Node.js 22.23.3 via fnm (upgraded from Node 20 because EAS CLI now requires Node 22+)
  5. Verified DNS records:
    • TXT _cf-custom-hostname.taze.to → 1266fffd-f4da-4b47-b167-4053ffaf5de6
    • CNAME _acme-challenge.taze.to → taze.to.7dede208f6a1d738.dcv.cloudflare.com
    • A taze.to → 172.66.0.241 (currently DNS only for validation)
  6. Removed Cloudflare Worker custom domains on taze.to and www.taze.to so EAS can take over
  7. Checked status.expo.dev — all systems operational

What I suspect

This looks like the known bug in EAS CLI where assignAlias doesn't verify the alias actually points to the new deployment, and the server mutation fails without a clear client-side error. Reference: PR #4443 in expo/eas-cli.

What I'm also requesting

Please clean up my account structure. I want:

  • Keep: lgo-studio (the paid account that should be billing and owning everything)
  • Keep: lg0144 (project owner)
  • Remove/unlink: lokoloko-2 (should no longer be involved with billing or this project)

Specifically, I want it confirmed that:

  • Billing is tied only to lgo-studio
  • The app-bilddin project is owned only by lg0144
  • lokoloko-2 has no ownership, billing, or access rights to this project

What I need from you

  1. Investigate why deploy --prod fails with a server error and let me know if there is a project-side issue I can fix
  2. If it's a server-side bug, please fix or escalate
  3. Confirm the account cleanup above (billing only on lgo-studio, lokoloko-2 removed)
  4. Once resolved, help me get taze.to live on the production deployment

Thank you.

Davy van Praet
lgstudio144@gmail.com

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    needs reviewIssue is ready to be reviewed by a maintainer

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions