What is your overall goal or next-step here? I think @grove’s gone over the options available to you, but I’ll repeat some of them as choices:
If you need to stand up a new GitLab instance, then you can check if the backups
directory you’ve salvaged carry any recent backups and perform a direct restore on a new installation. The backup carries your former database data in it, so without one you may not be able to easily create a restored instance.
If you’re merely looking to extract or to check the git repository information, you’ll need to clone each of their storage directories into working copies (as @grove mentioned, they are in a bare form in your copy of the git server’s directory):
~> mkdir /tmp/working-copies/
~> cd /tmp/working-copies/
~> git clone file:///path/to/copied/repositories/group/project.git project-working-copy
~> cd project-working-copy/
~> ls
Note: You’ve mentioned that you’re able to see the project structure under the repositories
directory, so I assume you were still on a GitLab version that used legacy storage. The new versions use a hashed path storage which will carry undecipherable paths that’s difficult to map to a nicer project and group name without the database backup available. You can still perform the above steps on their .git directories, but you’ll need to use git log
and other commands to manually work out which project they may have belonged to.