Git LFS "LFS: Fatal error: Server error" (500) in CI

By Kaveh Alemi·Latchkey

The LFS server or its storage backend returned a 5xx while transferring an object. Unlike a 403 or 404, this is a server-side fault and is often transient, clearing on retry.

What this error means

git lfs pull or checkout fails with "LFS: Fatal error: Server error: https://.../objects/..." or "Error downloading object: ... Server error (500)".

Error downloading object: assets/video.mp4 (7c6b5a4):
  LFS: Fatal error: Server error: https://github-cloud.githubusercontent.com/.../7c6b5a4
error: failed to fetch some objects

Common causes

A transient fault in the LFS storage backend

The object store hit a temporary error while streaming the blob. The request was valid; the server failed to complete it.

A large object timing out on the backend

Very large objects over a slow path can trip a backend timeout that surfaces as a 5xx to the client.

How to fix it

Retry the fetch

1. Re-run git lfs pull; a transient 5xx usually succeeds on a second attempt.
2. Wrap the fetch in a bounded retry so a single blip does not fail the job.
3. If it persists for the same object, check the provider status page.

for i in 1 2 3; do git lfs pull && break; sleep 10; done

Reduce concurrency for large objects

Lower the transfer concurrency so big objects are less likely to trip a backend timeout.

git config lfs.concurrenttransfers 2
git lfs pull

How to prevent it

• Add a bounded retry around LFS fetches in CI.
• Cache .git/lfs so a flaky server is hit less often.
• Lower lfs.concurrenttransfers for very large objects.

Frequently asked questions

Related guides

References

• Git reference manual
• GitHub Actions documentation