Red Hat SummitRed Hat blog
Red Hat OpenShift can run and manage both Linux and Windows containers from one platform, so you're able to move Windows containers to OpenShift without needing to completely re-architect or write new code. And this week, Christian Hernandez, Technical Marketing Manager for Red Hat Cloud Platforms Business Unit, joins the live stream to cover the ins and outs of bringing your own Windows nodes to OpenShift.
This stream we discussed the role and use cases for Windows nodes, and we explored configuring a cluster with hybrid Open Virtual Networking (OVN) Kubernetes networking and using the Red Hat Windows Machine Config Operator (WMCO) to install and manage Windows nodes.
We also spent a few minutes at the start of this stream to celebrate the one year anniversary of the Ask an OpenShift Admin live stream! 42 episodes in 52 weeks with more than 40 thousand views! You can hear even more statistics about the stream, and individual episodes, during stream at the time linked above.
Thank you so much to our audience! You are the reason we’re here and we hope you continue to find the stream valuable - don’t hesitate to contact us with requests and suggestions!
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 43 recorded stream:
Use this link to jump directly to where we start talking about today’s topic.
This week’s top of mind topics:
- This week we talked about several bugs and upgrade blockers for OpenShift. Including three which have been preventing OpenShift 4.7 to 4.8 updates.
- The first one is a kernel bug which caused CephFS corruption under some conditions. The good news is that this is fixed in 4.8.11 and 4.7.30!
- The second issue is a bug resulting in CRI-O having a runfs leak, where it does not free space after it’s done being used. This was fixed in 4.8.10 and 4.7.29.
- Finally, the last issue we discussed is an issue with vmxnet3 offload drivers resulting in packet loss. You can find more information about the problem, and suggested workarounds, in this KCS. It has also been fixed in OpenShift versions equal to or greater than 4.8.9 and 4.7.28.
Questions answered and topics discussed during the stream:
- Windows nodes have been available in OpenShift since OpenShift 4.6 and we talked about them previously with OpenShift 4.7 when vSphere IPI support was added. Today, we’re discussing the Windows Machine Config Operator (WMCO) version 3, which will add support for “bring your own host” when released in the near future. This means that you can add Windows nodes to any non-integrated or UPI deployment!
- What are the requirements of Windows nodes? Have there been any changes with WMCO version 3? No changes here, this release is focused on increased flexibility for adding nodes to the cluster.
- The prerequisites continue to be very important, in particular the need for OVN-Kubernetes with hybrid overlay networking enabled! The docs go in depth on how to deploy and configure the cluster for Windows nodes.
- Is there any harm or other concerns if the hybrid overlay is enabled at install time, even if you have no intention of using it? No, it’s present but unused until needed.
- A viewer asked us to elaborate on the details of the hybrid network. What does it do and what does the configuration mean? The short version is that hybrid allows the Windows and Linux nodes to communicate using different overlay protocols.
- Christian does a demo of adding a Windows node, starting here, where he walks through configuring the Windows operating system, the WMCO, and joining the node to the cluster.
