Ask an OpenShift Admin Office Hour - Operator Lifecycle Manager Deepdive

What's the best way to install, manage and upgrade Operators and their dependencies in a cluster? Operator Lifecycle Manager (OLM) is a core component of OpenShift. OLM is the Kubernetes feature/function that enables the deployment, management, and lifecycle management of all the Operators that make up OpenShift itself and the Operator-managed applications deployed.

In this stream, we take an in-depth look at the components of OLM and how they work together to manage everything from cluster deployment and upgrades to on-demand databases and services.

If you want an overview of Operators and their role in OpenShift, episode 9 of the Ask an OpenShift Admin livestream is a great place to start.

As always, please see the list below for additional links to specific topics, questions, and supporting materials for the episode!

If you’re interested in more streaming content, please subscribe to the Red Hat livestreaming calendar to see the upcoming episode topics and to receive any schedule changes. If you have questions or topic suggestions for the Ask an OpenShift Admin Office Hour, please contact us via Discord, Twitter, or come join us live, Wednesdays at 11am EDT / 1500 UTC, on YouTube and Twitch.

Episode 41 recorded stream:

Use this link to jump directly to where we start talking about today’s topic. 

This week’s top of mind topics:

• First up is a discussion about the current status of OpenShift z-stream releases, including a blocking bug for 4.7 to 4.8 stable upgrade edges. It’s been a few weeks since the 4.8 release, and we understand that not having a stable update edge is frustrating, but it’s reassuring that the process is working. This means that breaking bugs are being found and fixed before they have widespread impact!
• Which storage protocol should you use? NFS? iSCSI? Fibre Channel? Use whichever one works best for you and your application!
• If you want to add a second network adapter to your Machine API managed nodes the process is straightforward. There are several options for configuring the IP, including DHCP and static IPs using kernel parameters or the NMstate Operator.
• Finally, we mentioned that Nutanix has released blueprints to deploy OpenShift using their Calm automation tool. If you’re unfamiliar with this, Red Hat and Nutanix recently announced a strategic partnership to deliver OpenShift on Nutanix AOS. You can see a video of the deployment process here.

Questions answered and topics discussed during the stream:

• Every few days my cluster starts having issues as a result of the openshift-api and authentication Operators becoming non-functional. What can I do to troubleshoot and fix? Check the logs for OLM, the Operator, and the operand pods. There could be any number of things leading to this behavior, so it’s important to start by checking the logs.
• Does the VMware storage provisioner work with vSphere datastore clusters? Unfortunately, it does not. We go into some detail on options, alternatives, and gotchas with the provisioner when attempting to workaround this limitation, so be sure to give this 5 minute segment a watch if you have time.
• What is the role of Operator Lifecycle Manager? It deploys, updates, and manages Operators themselves, including the catalog of available Operators.
• There’s several excellent resources for learning more about Operator Lifecycle Manager in the documentation, both upstream and in Red Hat’s docs.
• What’s the difference between an Operator and an Operand? An Operand is an instance of the application which the Operator is responsible for.
• Troubleshooting tips for Operator errors.
• If you aren’t familiar with OLM, we walk through several of the components during the stream starting here, including the catalog, package server, and OLM itself. We also explore the components of an Operator from OLM’s perspective, such as the Cluster Service Version (CSV), and how to check on the deployment progress for an Operator using the InstallPlan object.
• How can you tell if an Operator is managed by OLM or has been installed separately? Look for a subscription object. The subscription for an Operator is what tells OLM that you want it deployed.
• As promised, here are the blog posts, part one and part two, for information about Helm vs Operators.