Migrate GitLab to GitHub? How to do it in an efficient and data-consistent way
Last Updated on June 14, 2024
“It depends.” – probably the most favorite phrase of programmers and lawyers. “It depends” because there is always a hidden context behind any question or problem. In addition, everything is changeable, so the solution always relies on something. In the IT industry, which is probably the fastest-changing industry in the market, this is clearly visible. Many solutions or technologies have existed in projects for many years. Some are already obsolete, but this is the result of past decisions that were probably right at the time, and yet the environment has changed.
When various application servers were created, did anyone think about containers or orchestration at the time? Of course not, because projects like Docker didn’t exist then. The same was with version control systems and source code repositories. Most IT projects today use various hosting services of the Git system, but it was different back then. The SVN system used to reign supreme and, in the context of the time, it was a good choice. It was only the passage of time and the development of other technologies that led to the necessity (or possibility) of migrating projects from SVN to Git. The Git version control system, being a distributed system, allows developers to save snapshots of the project locally, commit changes without connecting to a central server, establish branching and merging, and track project history. However, keep in mind that Git itself is “only” an engine, a tool that requires hosting – either own or external. The latter is the most common choice, and the most popular hosting services have tens of millions of users each.
Which Git hosting service should I choose for my Git repository?
Currently, there are three most popular services. They are Bitbucket, GitLab, and GitHub. At first glance, they do the same thing, though they differ in detail. GitHub strongly supports open-source projects, and in recent years, GitHub Actions has been betting on the possibilities offered by CI/CD. GitHub now allows users to create unlimited private repositories for free, which was previously a feature more commonly associated with GitLab. GitLab, which seems to be a step ahead of GitHub when it comes to the aforementioned automation, also offers the ability to self-host GitLab on our own servers. Bitbucket, on the other hand, as a tool from Atlassian, integrates very well with its other applications, like Jira, Confluence or Bamboo.
Read more about the different hosting services and the differences between them here: TOP Git Hosting services 2023.
I mentioned that GitLab might be a better solution from a CI/CD perspective. Well… maybe. But that’s not the only determinant. GitHub is certainly an equally good tool, on top of that, it is being developed very rapidly. Also, what is worth mentioning is the GitHub Enterprise Server service and its Backup Utilities, which is a backup system installed on a separate host. It works as a typical backup tool, by making regular snapshots of the original server. However, we can also use third-party solutions to do such work in a more convenient way.
GitLab previously allowed users to host their own private repositories for free, while GitHub only allowed private repos for paying customers. However, GitHub some time ago announced free unlimited private repositories for individual users, which has led to users considering transfer repositories or migrating their projects to GitHub.
To find more about other options check our GitHub CI/CD guide for beginners.
Cross-over recovery
Speaking of third-party tools, there is another benefit they can provide us with. Especially on the topic of migration. GitProtect.io has a feature called cross-over recovery, that is, the ability to backup the entire GitLab environment and restore it to another service provider – in this case, it’s GitHub. You can add multiple organizations and different accounts – there is no limit. The biggest benefit of this solution we can see in the case of multiple repositories. And by “multiple” I mean hundreds or thousands. Cross-over recovery is enabled between GitHub, Bitbucket, and GitLab in any direction.
Are you switching to a DevSecOps operation model? Remember to secure your code with the first professional GitHub, Bitbucket, GitLab, and Jira backup.
Simple as that! Just select what you want to copy. This tool will make a copy of repositories and metadata. Then just pick a new or existing GitHub organization as a data recovery destination. If you are transferring from an external git repository, you can create a bare clone and push it to GitHub. And that’s it. GitLab to GitHub migration can be considered complete with all references copied to the imported repository. For more information about this tool, I proceed to the official site.
Using this functionality, the migration process itself is simple, and basically boils down to a few clicks on the graphical UI in GitProtect.io. Each backup has a “Restore” function, visible on the right side of the screen:
Next, we need to select the destination where we want to restore our backup. Most often it will be the same service, local machine, because we will usually want to perform a restoration plan in case of a disaster, for example. However, we can use this option to migrate and restore to another service:
As you can see, it takes just a few simple steps to migrate GitLab to GitHub. Of course, at the backup stage, we also have control over the metadata that this copy will contain. At the time of migration, we will also be able to choose which of this metadata we want to migrate.
Step-by-step process to import repository
For comparison, I will describe how this can be done without using external tools. Granted, this is not rocket science, however, you will see for yourself that the above steps are much easier and faster. Let’s find out for ourselves.
The entire migration process consists of several stages:
- preparation – here we need to ensure that our repository is as up-to-date as possible, so all changes, tags and metadata should be integrated into our repository
- cloning – namely, creating a copy of our Git repository. It is crucial that such a copy contains all tags and branches. During this step, you should also add an external clone URL for importing the repository to GitHub.
- saving – the finished, downloaded copy must be saved to the new, target repository address to which we are moving our project. This involves creating a temporary local repository, pushing it to the new repository, and then removing the temporary local repository after the import process is completed.
- configuration and sharing of the repository – here it is worth checking or correcting any security configurations, pipelines, accesses, etc. Ensure that the locally cloned repository is correctly pushed to GitHub using the ‘mirror’ option.
The greatest risk of an error and workload lies in the first and last steps. Human errors can cause even the best and the most accurate plans to fail. That’s why process automation is so important, as it minimizes the risk of such errors and makes the process of GitLab repository transfer to GitHub repositroty smoothly.
Boiling down the above plan to Git commands, it can be written this way:
- git clone –mirror <GitLab_repo_address>
- git remote rm origin
- git remote add origin <GitHub_repo_address>
- git push –mirror
Conclusion
There are a few ways to migrate GitLab to GitHub, which method works for you is absolutely your choice. Hence, you should remember that there may be many reasons to migrate GitLab to GitHub. Anyway, as always, we need to be prepared for unexpected situations or errors and always have a backup and recovery plan on standby. Using third-party tools, like GitProtect.io, not only gives us this but also it provides the possibility of cross-over recovery. All of that not only keeps us safe but also makes the GitLab to GitHub migration process very easy.
When migrating, configuring credentials using a GitHub personal access token is essential for a smooth transition. This token helps in setting up the required credentials for various tools and commands during the migration.
Additionally, third-party tools can help migrate metadata such as issues, milestones, and pull requests, ensuring a comprehensive transfer of your project data.
