To swiftly reboot one’s IT infrastructure, prior automation is essential. This fundamentally alters the approach to administration. Embarking on the journey towards data center cold-start capability involves replicating a real-world environment that has been tried and implemented to a large extent as described. This stems from a multitude of insights and decisions made as part of a long-term strategy. One of these insights was to automate IT operations, ensuring a more sustainable use of administrators’ knowledge and time.

We manage a medium-sized environment with a moderate four-digit number of network components, around a thousand virtual machines, and nearly one hundred hosts for virtualizing various sizes alongside a blend of software-defined storage (SDS) employing diverse technologies.

The path to establishing a cold-start capable data center infrastructure is lengthy and involves intensive learning processes. The foundation of cold-start capability lies in automating deployment and maintenance to swiftly reboot the cold-start capable infrastructure.

Every IT automation relies on three pillars: inventory for assets and their attributes, repository storing and maintaining actions, and orchestration executing changes on assets. Within our setup, NetBox serves as the inventory or Configuration Management Database, GitLab functions as the repository, and Ansible handles configuration management.

Understanding storage systems and the maturation of operational processes have steadily grown in recent years, leading us to prefer different systems than in previous years.