![]() ![]() Our long-term goal is to move all projects off the old instance to the new one but as we, as site reliability engineers, know: permanently turning off a service is harder than one thinks. ![]() The newer instance is completely automated regarding user/repo management while the older one is a chaotic mess of manually created repositories, groups and assigned permissions. The other one is newer and contains the majority of our active projects. One of these is older and was set up at a time when we had comparatively few projects and employees. Current Architecture and migration targetsĪt the start of this journey, we were operating two self-hosted GitLab instances. Utilize hosted offerings whenever it makes senseĪdditionally, we can use this opportunity to pick up some new possibilities along the way like zero-downtime upgrades. No difference should be noticeable for end users (this means changing URLs/ports is a no-go)Ĭost-Effective resource sharing with other services Only allow access from our company network Requirements and Constraintsīefore designing the target architecture, we have to analyze the requirements and constraints. It should not serve as a guide, but help you on your journey moving GitLab to the cloud. This blog post outlines the path we took, the decisions we made and the challenges we encountered while moving these GitLab instances to a Kubernetes-based setup. ![]() Most of these services were already running on OpenShift which made the migration pretty painless and straightforward but our GitLab instances were running on virtual machines to avoid complicated networking setups concerning SSH in OpenShift. This includes GitLab, TeamCity, Nexus and a lot of other supporting services, enabling us to build software the way we do. At Cloudflight, we recently moved our on-premise build infrastructure to Azure. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |