Skip to content

FAQ

Jump to bottom
Csaba Polyak edited this page Sep 28, 2026 · 1 revision

FAQ

Does this back up private repositories?

Yes.

The GitHub API is authenticated with a Personal Access Token, so repositories accessible to that GitHub account can be discovered, including private repositories.

The SSH key used for Git operations must also have access to the repositories being mirrored.

Why does the backup need both a token and an SSH key?

They are used for different purposes.

Personal Access Token

Used for:

  • repository discovery
  • GitHub API authentication

SSH key

Used for:

  • cloning repositories
  • updating existing mirrors

The token is never embedded into Git repository URLs.

Does this require GitHub CLI (gh)?

No.

The script uses standard tools:

bash
git
curl
jq
ssh

GitHub CLI is not required.

Does it back up all my GitHub repositories?

It backs up repositories returned by the GitHub API for the authenticated account.

This includes repositories where the account is an:

  • owner
  • collaborator
  • organization member

Repositories to which the account has no access cannot be discovered or mirrored.

What happens when a repository is deleted from GitHub?

The local mirror is not automatically deleted.

Instead, the script reports it as an orphaned mirror.

ORPHANED MIRRORS

! project-old.git

The backup is intentionally preserved.

This prevents an accidental GitHub deletion from immediately becoming a local backup deletion as well.

What happens if a repository is renamed?

The old local mirror is preserved.

Depending on how the repository appears during discovery, it may subsequently be treated as an orphaned mirror while the renamed repository receives its own local mirror.

This means a rename does not automatically destroy the old backup.

What happens if one repository fails?

The backup continues.

A failed repository is recorded and the remaining repositories are processed.

The final exit code becomes:

1

instead of 0.

This allows scheduled monitoring to detect the problem without losing the rest of the backup run.

What happens if the backup is started twice?

The script uses a lock directory to prevent concurrent executions.

The second instance detects the existing lock and exits.

This prevents two processes from modifying the same mirror simultaneously.

Is the first backup slow?

It can be.

The first run has to download every repository and all required Git objects.

The amount of data depends on:

  • repository count
  • repository history
  • repository size
  • network speed
  • storage performance

Later runs are incremental and normally require considerably less data transfer.

Does it back up GitHub Issues and Pull Requests?

No.

The current project is a Git repository mirror, not a complete GitHub account backup.

GitHub-specific metadata such as Issues, Pull Requests, Actions artifacts, Releases metadata, permissions and settings require separate backup mechanisms.

Does it back up GitHub Wiki?

Not as part of the normal repository mirror.

A GitHub Wiki is technically a separate Git repository and requires separate handling if it needs to be included in the backup.

Does it back up Git LFS?

Not automatically as a complete LFS backup.

Git LFS stores large objects separately from normal Git objects.

Repositories using LFS therefore require an additional LFS backup procedure.

See:

Git LFS

Can the mirror be restored to GitHub?

Yes.

Create an empty destination repository and push the mirror:

git --git-dir=/path/to/repository.git \
    push --mirror git@github.com:USERNAME/repository.git

See:

[Recovery](Recovery)

Can the mirror be restored to another Git server?

Yes.

The mirror is a normal bare Git repository and is not tied to GitHub.

It can be pushed to another Git server supporting Git over SSH.

Does the script require Docker?

No.

The script is intentionally designed to run directly on the host.

This makes it suitable for NAS systems and small Linux servers where running an additional container would provide little benefit.

Does the script delete old backups?

It does not automatically delete orphaned repository mirrors.

The only automatic cleanup currently performed is log rotation.

Old log files beyond LOG_KEEP are removed.

Repository data is intentionally left untouched.

Is the GitHub token stored in the repository?

No.

The token is read from a separate local file:

~/.config/github/token

It should never be committed to Git.

The script also avoids placing the token into Git remote URLs.

Can I use it on a Synology NAS?

Yes.

The script was designed with NAS environments in mind.

It can be executed manually or through Synology DSM Task Scheduler.

See:

[Installation](Installation)

and

[Scheduling](Scheduling)

Do I need to keep the GitHub account logged in?

No.

The backup is non-interactive.

The GitHub API uses the stored token and Git uses the configured SSH key.

This allows the script to run unattended through cron or a NAS scheduler.

What happens if GitHub is temporarily unavailable?

The API and Git operations can fail independently.

The script reports the affected operation and continues processing where possible.

Individual repository failures result in exit code 1.

A fatal failure that prevents the backup engine from operating results in exit code 2.

Is this a complete disaster-recovery solution?

It is a Git repository backup tool, not a complete GitHub disaster-recovery system.

For source repositories, a mirror provides a strong independent copy of the Git data.

A complete GitHub account backup would additionally require handling GitHub-specific metadata and services.

The scope is intentionally kept small:

GitHub
  ↓
Git repositories
  ↓
Local mirrors

Simple enough to understand, automate and recover from.

Clone this wiki locally