{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiegwjk2occdkcupujeoxnnc7farmlvmo34ty2en77tcnvensyf37a",
"uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mhxjoqbw63j2"
},
"path": "/t/docker-space-build-stuck-empty-logs/174644#post_1",
"publishedAt": "2026-03-26T10:44:31.000Z",
"site": "https://discuss.huggingface.co",
"textContent": "**Docker Space Build Stuck — Empty Logs**\n\nHi, I’m experiencing an issue where Docker Space builds are permanently stuck in `BUILDING` with completely empty build logs since today (March 26). The same setup was working fine yesterday.\n\n**What happens:**\n\n * After pushing to a Docker Space, status shows `BUILDING` but build logs are completely empty — nothing appears at all.\n\n * The API shows `hardware.current: null` while `hardware.requested: cpu-basic`, so it seems no build machine is ever allocated.\n\n * Factory reboot has no effect.\n\n\n\n\n**What I’ve tried:**\n\n * Deleting and recreating the Space multiple times\n\n * Creating a minimal test Space with just a 2-line Dockerfile (`FROM python:3.11-slim` + `pip install streamlit`) — same issue\n\n * Different Space names — same result\n\n * Token is valid and has `repo.write` permission (same token worked yesterday)\n\n\n\n\n**Additional issue:** After deleting and recreating a Space with the same name, the API returns `Invalid input: expected object, received undefined → at metadata/compute/service`. This persists even on a freshly created Space with that reused name, suggesting the name gets corrupted in the backend after delete/recreate cycles.\n\n**Account:** `yc-lin-geo`\n**Current affected Space:** `yc-lin-geo/ISMIP6-Antarctica-SLR-Explorer`\n**Last successful build:** March 25, 2026",
"title": "Docker Space Build Stuck — Empty Logs"
}