Is Running Oracle Database on Kubernetes a Viable Option Nowadays?

You can run Oracle on K8s, but expect a complicated setup and maintenance. If your customer insists on it, ensure you’ve discussed the implications of Oracle's licensing policies. It's often recommended to consider alternatives if possible, as running Oracle in K8s can lead to more headaches than it's worth.

The old wisdom of avoiding databases on K8s seems to be fading. Modern setups with Operators have transformed the landscape, but Oracle presents unique challenges with its licensing. If you really have to use Oracle due to customer demands, consider isolating it on dedicated nodes to simplify the licensing issues. Just be aware that this can add extra complexity to management!

You definitely need to be cautious about licensing. Oracle has strict requirements, and they usually want you to license every core in your entire cluster since K8s pods can be scheduled anywhere. They don’t really acknowledge the intricacies of namespaces or node affinities as valid partitioning. You might end up needing to set up isolated clusters for Oracle, which kind of defeats the purpose of using K8s in the first place.

Running databases in K8s is definitely doable with the right setup! Operators and CSI drivers can make the management much easier than before. However, you still need to be quite experienced because databases can be tricky to manage, and mistakes can lead to data loss. Make sure you have a solid recovery plan and understand K8s well before jumping in.

I’m not a fan of running Oracle on K8s. There are a lot of licensing and performance issues involved. If you can, look into alternatives like PostgreSQL, which can run smoothly in K8s. It reduces the complexity and headaches associated with Oracle's policies.