Managing your cloud ecosystems: Maintaining workload continuity during worker node upgrades

Planning and managing your cloud ecosystem and environments is crucial for decreasing manufacturing downtime and sustaining a functioning workload. Within the “Managing your cloud ecosystems” weblog sequence, we cowl totally different methods for guaranteeing that your setup capabilities easily with minimal downtime.

To start out issues off, the primary subject on this weblog sequence is guaranteeing workload continuity throughout employee node upgrades.

What are employee node upgrades?

Employee node upgrades apply essential safety updates and patches and must be accomplished repeatedly. For extra info on sorts of employee node upgrades, see Updating VPC worker nodes and Updating Classic worker nodes within the IBM Cloud Kubernetes Service documentation.

Throughout an improve, a few of your employee nodes might change into unavailable. It’s essential to ensure your cluster has sufficient capability to proceed operating your workload all through the improve course of. Constructing a pipeline to replace your employee nodes with out inflicting software downtime will assist you to simply apply employee node upgrades repeatedly.

For traditional employee nodes

Create a Kubernetes configmap that defines the utmost variety of employee nodes that may be unavailable at a time, together with throughout an improve. The utmost worth is specified as a share. You can even use labels to use totally different guidelines to totally different employee nodes. For full directions, see Updating Classic worker nodes in the CLI with a configmap within the Kubernetes service documentation. When you select to not create a configmap, the default most quantity of employee nodes that change into unavailable is 20%.

When you want your whole variety of employee nodes to stay up and operating, use the ibmcloud ks worker-pool resize command to briefly add additional employee nodes to your cluster all through the improve course of. When the improve is full, use the identical command to take away the extra employee nodes and return your employee pool to its earlier measurement.

For VPC employee nodes

VPC employee nodes are changed by eradicating the previous employee node and provisioning a brand new employee node that runs on the new model. You may improve a number of employee nodes on the similar time, however when you improve a number of without delay, they change into unavailable on the similar time. To be sure to have sufficient capability to run your workload in the course of the improve, you’ll be able to plan to both resize your worker pools to briefly add additional employee nodes (much like the method described for traditional employee nodes) or plan to improve your employee nodes one after the other.

Wrap up

Whether or not you select to implement a configmap, resize your employee pool or improve parts one-by-one, making a workload continuity plan earlier than you improve your employee nodes will help you create a extra streamlined, environment friendly setup with restricted downtime.

Now that you’ve got a plan to stop disruptions throughout employee node upgrades, hold an eye fixed out for the following weblog in our sequence, which is able to focus on how, when and why to implement main, minor or patch upgrades to your clusters and employee nodes.

Learn more about IBM Cloud Kubernetes Service clusters