Ensure documentation is sufficient for NFS migration to Gitaly Cluster

Problem to Solve

With the deprecation of NFS support coming in %14.0, many self-managed users are interested in migrating to Gitaly Cluster. This is the preferred path, so we want to ensure there is a clean pathway for these users.

This is also a one-time problem for repositories, as once the migration is complete, they won't need to migrate the repository again. Therefore, we want to balance ease of use, with the appropriate engineering investment given it's limited lifespan.

Potential Solution (MVC)

There exist ways to migrate repositories today, but they involve performing a series of manual steps. While this may be perfectly sufficient for our needs, we need to ensure that there is clear documentation.

As an MVC, I would propose to update our documentation to outline the commands / operations necessary to migrate a repository from Gitaly NFS Storage to Gitaly Cluster. I would also like to recognize that many organization who self-host have numerous repositories, so we want to make sure that the solution lends itself to scripting or batching in some capacity. It's not reasonable to expect a user to migrate 1000 repositories one at a time by hand.

Acceptance criteria

• Ensuring customer data is a paramount priority. We want to make sure we offer a non-destructive way of migrating data so that if problems occur, there is no loss of data.
• We need to ensure that the steps can be followed by a reasonably experience Systems Administrator (our Sidney persona)
• We should strive to have someone outside of the Gitaly team test the steps and provide feedback (perhaps someone in the SA organization).

Timeframe

Since the deprecation is taking place in %14.0, I would like to get this completed within the %13.11 milestone to ensure there is adequate time for our users to migrate if they choose to do so.