Red Hat OpenShift Ask an OpenShift Admin | Red Hat Livestreaming, 7 Dec 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Ask an OpenShift Admin (E89) | Dive into OpenShift Virtualization

Description

OpenShift Virtualization is a feature of OpenShift, when deployed on bare metal servers, to host virtual machines alongside containers in the same cluster. They can connect to the SDN, consume PVCs, and do many of the other things that Pods do - but also have VM behavior like live migration.

Ian Pilcher joins us this week to show some capabilities of OpenShift Virtualization and how it might fit into your application modernization strategy.

Want to learn more?

Try Red Hat OpenShift: https://red.ht/3uVxkya

Follow us:
Andrew Sullivan https://twitter.com/practicalandrew
Jonathan Rickard https://twitter.com/jrocktx1
OpenShift https://twitter.com/openshift

Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_confirmation=1

About Ask an OpenShift Admin:
Do you have questions? We want to answer them! Every Wednesday at 11am ET, hosts Andrew Sullivan and Jonathan Rickard sit down to answer your questions about Red Hat OpenShift.

About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3rYz5ZS
OpenShift on GitHub https://github.com/openshift/

#RedHat #OpenShift #Kubernetes #ContainerPlatform

A

A

Let's get into the big topic of Open Source.

B

C

Hello and welcome to this week's ask an openshift admin office hours live stream. I am Andrew, Sullivan joined, as always by my amazing and awesome co-host, Mr, Johnny, Ricard, happy Wednesday.

A

Sir happy Wednesday to you too man, yeah.

C

It's good to good to be here. I feel, like we've been gone for a Month of Sundays, but uh it's it's only been like what three two weeks.

C

Three weeks we hadn't we had an accidental uh cancel last week because it was a combination of you and I were taken by surprise and we we thought we had a topic and then we didn't have a topic, and then we thought we had a guess and we didn't have a guest, and so we just decided to the other than completely wing it we'll just uh we'll cancel so right, somewhat professional.

A

Right it was, it was the right move, yeah.

C

So uh you know apologies to the audience. um You know for anybody who is exciting or waiting. um uh You know it's it's good to be back, though, and uh I am very happy to be joined today by one of my peers, uh so Mr Ian Pilcher, welcome to the uh streamian.

B

Hey. Thank you. Thank you very much. I'm I'm just here to try to live up to your high logistical standards.

C

Yeah, you know we're we're fancy here right, it's nothing, nothing but the as Johnny said right, highest standards, highest standards, uh so, as you can tell by the topic or the title of today's stream, rather we're going to be talking about openshift, virtualization and uh Ian is an expert on many things and has been an expert on many things.

C

Virtualization for far longer than I have known you Ian, um you know, you've, you've done a lot with uh openstack and like KVM and Rel and all kinds of stuff and openshift virtualization is just the the newest one in the line. If you will so.

B

Yeah there's a reason: my.

C

Head's so shiny I'm, I I can't comment on anything. I got. You know way way more gray,.

A

C

You know uh Johnny's growing his beard out he's starting to look a little bit like a mountain man. There that's.

A

Right, that's what happens when you let it go. You know once you just let go of like you know everything that looks wise. This is what this is.

C

So yeah and we were chatting just a minute ago- um you know today I'll say we don't have a super strict like normally, you know we'll go through and we'll create an outline of all the things we want to talk about and show and I know. Ian has a few things he want to show, but um really for the audience like I Ian's, super technical I and again, let's, let's, uh let's go in the weeds like let's see what what what's happening underneath the covers. You know I'm, I'm, I, don't want to challenge anybody.

C

um What's that uh uh stump the chump right I, don't want to play stump the chump necessarily, but uh you know it's been a while, since we've talked about openshift virtualization and um without wanting or or trying to um you know, incur the raft so to speak, you know there's a lot of noise going on on the virtualization World these days. um So you know, let's have a conversation.

C

Let's look at a few things that are going around and you know to the audience feel free to ask questions uh at any time, uh speaking of which, so it has been a few weeks and I feel like I. I often forget to do this reminder of this. Is an office hours live stream? uh So what that means is that we are here to answer questions about anything and everything open, openshift related. If I could talk, uh that's on your mind.

C

So, yes, today we have Ian with us we're talking about openshift virtualization, but don't let that limit the questions that you are that you want to ask about. So anything, that's on your mind. Any questions, problems, issues, concerns valid or invalid that you may have we're happy to talk about those um yeah, Quay 3.8 got released. I think you are hope. Nine I think that's in our top of Mind topics to talk about in just a moment, um so yeah um again keep those questions flowing in. um We love to chat.

C

We love to hear from everybody so uh and with that we will roll into top of mind and go from there. uh So Johnny honestly I just glanced at the list. I remember I was throwing things at you in slack this morning, but I don't remember what all of them were.

A

Yeah I think it was Clay 3.8 right, I mean uh so clay, 308 I think released yesterday or at least this week, and um it looks like there's a lot of exciting things that you know um that are going to be coming out with it. You know like there's a new UI, there's going to be some added uh restrictions or not I shouldn't, say restrictions, but added capabilities to to create restrictions within Quay and then um better container image management.

A

You know so um I think that the the one thing that that kind of caught my attention was that the storage size is going to go from base two to base 10 or maybe I might have that inverse. It might be base 10 to base two. It's going to go from, maybe bytes to megabytes yeah, so two to ten and so it'll make that, like when you're looking at your image sizes it'll make it look a little bit more.

A

um You know normal I. Guess is one way to put it, but then yeah so like when you, when you're looking at like your existing play and then like this upcoming play, that you might see some differences there, because that that base of the storage is going to change.

C

Seems like a weird direction to go going from two to ten.

A

Yeah because I thought everything was in like the mibian Gibby and whatever the petabyte one is maybe, if I'd give.

B

You, the storage, is often quoted in the base 10 stuff, so.

C

Yeah yeah: well, the marketing people do anyways.

A

C

A

All know, storage guys are mean so yeah, because.

C

I I can vouch for that. I was a storage guy for like a decade um we're the grumpiest people in the data center for a reason right.

B

You and the dbas yeah yeah.

C

So actually I I used to have a whole. You know I used to work for a storage. Vendor I had a whole rant about this. So if, like the reason why storage people are so grumpy is because of virtualization people and the the logic being way back when you know so think, 2008 2010, like almost overnight because of the global recession, storage arrays in the data center, went from just hosting application data to hosting every IOP of every server in the entire. You know Enterprise, and so they were woefully undersized to take on.

C

You know what in many cases were thousands or tens of thousands of VMS, so every performance problem for years it was a result of storage just being undersized, and then you know, storage companies got smart and started doing things. Like hybrid storage, you know flash accelerated, disk and eventually flash but yeah it went from uh you know it was for for years and years and years it was always the Network's guy's fault right. Oh, why isn't this working? Oh, there was a blip, oh yeah. We were about it. We rebooted this router.

C

The packets will just flow around it's fine and then it was the it was briefly the virtualization team's fault, and then it was the storage teams fault so anyways.

C

You know many many fond memories of having those conversations. uh You know Johnny, as many folks know, Johnny and I used to work together. We we probably I, would not be surprised if you and I had that argument with each other. At one point, John.

A

C

uh So yeah um interesting change, um I I, honestly haven't looked at what all is going on inside of clay I. um With this most recent release, I will be completely honest. I have fallen a little bit behind on keeping up with those things. um As uh both of you know, you know, I got tasked with doing a big project for IBM, so I was basically in a hole isolating from the rest of the world for about a month until we got through all that other stuff.

C

A

What's that it's a professional, coven, yeah.

C

Yeah well, but yeah I'm glad that's done now. So now we're just getting ready for uh the holiday season and all that other stuff, um let's see, moving on ACS 373 was released. um This is another one. I will admit that I have not kept up with Johnny I, don't know if you can speak to that.

A

Now the one thing that you called out was uh the migration from the internal database into a more public, or you know, uh or so consumer grade yeah database right I can't remember what it's I don't know. If it's going from roxdb.

C

Yeah yeah I I saw I was looking through the release notes, and it said it was going from the proprietary rocksdb I had no idea that there was a custom database being used inside of that thing, that's that's news to me: I'm happy to see we're moving away from it. um I will say you know sort of word of warning to folks. um So, if you're using ACS today when you go to 373 um I, think postgres becomes optional, but it for the next version It's required.

C

So you want to make that transition before you go to the next version, and uh I also want to say that the managed service side of things is already using postgres. So there's that as well.

C

um Our Hope nine is that because koi will now be able to set quotas per repo, uh I, don't know um I hope. So that would be good if we can yeah.

A

That's that's one of the benefits. That's one of the new features that are coming out in three eight is that you can set it you're gonna be able to set it per repository, there's some more stuff with like the super user, where um you know like now, the super user doesn't have to own the actual um repository where there's actual R back built in, and it looks like it's a lot more uh streamlined now than yeah than what it's been in previous releases. Yeah.

C

Yeah I know that the Quake team has put a lot of work into um modernizing their code base. Some folks may remember: it's been like six months ago. They had like a full day outage on cuer.io, as they flipped the code base to the new one, and there's been a lot of improvements since then, um Manish uh Manish apologies for butchering names. um Does this mean we can use openshift for creating virtualized environments and virtual machines? uh If true, can we replace VMware vsphere? So we will talk about that here in just a moment.

C

So um the the short version is yes, but no or no. But yes, maybe no.

B

But yes, depending on how you want.

C

To look at it yeah, so yes, absolutely openshift can deploy virtual machines, but it is not. We do not recommend it as a vsphere replacement for a number of different reasons. Okay,.

B

I'll I'll, let me give you my version of it um open shift. Virtualization is not a direct replacement for VMware openshift virtualization can replace VMware, and you have to stay tuned to explain why that actually makes sense.

C

B

C

Thank you for the uh teaser Ian.

B

C

uh Yeah matrushka VMS yep: it's a nesting dolls, all the way down turtles turtles, all the way down.

C

uh Let's see what else is in our list here: coupon sessions- oh, this is a good one, uh so the uh uh Alex over at the um blog sites, I'm gonna post this link into the chat here um so Alex handy, who runs the cloud.red hat.com blog uh collected all of the kubecon sessions that red Hatters participated in and put a link up there to all of them.

C

So if you're interested in any of the stuff that we had going on at kubecon, in addition to you know the booth and and all the press releases and everything else, I think we had. It was like dozens of press releases that week um it was a quite the media blitz, um but if you're interested in the actual technical stuff, all of the you know the real reason why people go to kubecon that links to all the stuff that uh that we did there.

C

So it's a good blog post, um I've perused through a couple of them so far and as you would expect from kubecon they're, always really good about selecting great sessions.

C

uh What's next Johnny I think we got.

A

C

A

Yeah, the the big one like really the the headliner is um Argo CD graduated to uh um within the cncf to it's a graduated project. Now so it's it's the real deal, the real project and um their official official as Christian put it.

A

um So congratulations to the Argo CD team and you know that's good news for us with openshift get Ops and um yeah openshift get Ops as a as an operator and as a product within red hat and openshift is just going to grow and get better and better and um 2023 has got a lot of good things on the roadmap. So I'm really excited to see what they, what they're able to get done.

C

Yeah yeah, you know who, who knew get UPS would be a thing and would be so popular. You know, oh right, a lot of full-time like Christian, uh you know yeah, it's uh I'm really happy to see that I graduated I think it reinforces the popularity and the power behind you know Argo and the get Ops, mentality and philosophy.

C

um You know case in points. We we have a peer, you know um live stream, you know, that's focused on get UPS, uh you know Johnny. Basically, your full-time job is get Ops uh so and I know you're, not you're, not the only one, not only in red hat but across um you know other other parts of the industry. So it's a great tool. It does a lot of really cool things. You know, we've talked about it briefly here.

C

um I, don't remember if it was the last episode or the one before that we talked about how get Ops applies for administrators and the ways that we can use it.

C

um You know it's being integrated and being taken advantage of in many other ways as well uh ACM you can use application sets directly from inside of ACM.

C

You know, which is the Argo application definition: um language object, crd, that's what I'm going to crd resources yeah, um so it's becoming more and more pervasive I would definitely suggest that you know we should all, as administrators do at least a little bit of due diligence and make sure that we're aware of it, because more than likely, your application teams are using it as well all right, um yeah Let's.

C

uh Let's talk some openshift virtualization, um so I I think uh Manisha has has set us up a little bit here um to start the conversation around openshift virtualization, so I I know, and you know, I have no idea which, what you have to demo or what you want to show or anything like that. um But two minutes uh question you know. Yes, absolutely we can deploy virtual machines um and the the fun part is the virtual machines in containers. um You know it's, it's still kubernetes. It's still openshift.

C

um You know we're not bypassing that aspect of it. So uh please I I, don't I don't want to steal your thunder.

B

Right yeah, this is this, is yeah I mean I. Think that's exactly the point. Is it's virtual machines running within openshift um as as VMS, but also as pods and being able to use all of the Technologies and operational practices that you use in an openshift environment and apply those to your virtualization environments?

B

um So with that, let me uh let me go ahead and share the screen here. I think we did promise these folks a live demo somewhere along the line. Hopefully that will uh come up here. Awesome awesome. So, as you can see, this is the uh this.

B

Is the openshift console that we all know and love, um but my open shift console in this case also has this virtualization menu item added um and this this just shows up when you install the openshift virtualization operator, um and that's that is like three Clicks in operator Hub um to do that So within the virtualization um Sub menu. We get this overview tab.

B

This will list, you know, give us an overview number of VMS. We got running. I got a big one here, as you can tell um also provides metrics, although it looks like I, don't have anything using any actual resources. So that's very unexciting right now, but once you have actual virtualized virtual machines running in your cluster, um you will see the uh that you know these various metrics of CPU utilization.

B

um Let's see Emery swap CPU weight, throughput iops Etc. We also have the virtual machines um that shows you virtual machines that have been created in the cluster. um So you can see here. I have a single virtual machine defined in this namespace um and there's machines like like everything else, they're namespaced and in fact, let's just go up here and we'll see that we have this container or this pod here, what we call a vert launcher pod.

B

So when a virtual machine is started, uh we actually create a pod called the invert launcher pod, um and that is where we run the virtual machine. So the virtual machine doesn't run as some special snowflake thing. It runs as a pod, um just like any other pod within your cluster.

C

Yeah and I I wanna, maybe I'm, jumping ahead here; sorry I'm, half choking because I'm cough having a coughing fit. um If you, if you look in that launcher pod, it's just a it's. A KVM virtual machine right, yeah.

B

And let's we can, in fact we can do that. Let's be uh let's be dynamic here. So.

B

B

Thank you, that's is it two dashes or one in the new syntax I. Don't remember.

B

All right, so there we are.

B

There we are all right, I knew I would get it eventually. So here we are, you see, we've got a liberty process running, we've got a long, D process running, um and then there is the virtual machine itself. You know, which is exactly what you would see. You know, on a on a normal Linux box, running a KVM virtual machine.

C

So I I assume I, don't know if the command is in the pilot, but you could do like a virtual and you know like a virtual list and see the VM there. Just like any other is.

B

There all right yeah there. It is okay,.

C

So yeah, so it's it's literally! It's KVM like this is what's.

B

C

Both uh amazing and baffling to me at the same time right because if you think about it, it makes a logical sense. You know what is a virtual machine. A virtual machine is a process and like what does a container? Do a container encapsulates a process so so long as the container grants access to the resources that the process needs.

C

It just works. It's like it's no different than any other process. um So yeah it's! It's always. It took me a moment when I first started working with openshift virtualization to really wrap my head around this concept of like it's, not it's not doing anything, really unnatural. It's just adding a layer of management around that virtual machine process.

B

Yeah, it's it. That's that's exactly right and in fact, in a lot of ways it's you know it can be a bit of a struggle to try to present this topic because it could be one slide it could, it could literally say you know, run run. Qem, run qem ukbm in a pot yeah next question: um yeah the the reality is there. There are a lot of details.

B

um We, you know, we expect virtual machines in traditional environments to be able to do things that uh we don't expect pods to be able to do in a kubernetes environment, and we've done a lot of work to bring a lot of those capabilities to openshift virtualization so that as people bring workloads from those traditional environments, they're not going to lose the the functionalities that they're used to um one of which you know is having a you know, an administrative interface that you know doesn't require you as an example to write 100 lines of gamble to provision a VM.

B

So yeah we'll go a lot quickly through the the Wizard approach here. But I do want to look at um what a running VM looks like inside the console. So let me check on you know, click on this one and pretty standard stuff. You got the tabs up here, giving various bits of information.

B

um Let's see, we do have a VNC console available. So if you know, if you lose network access or need to set up something, troubleshooting Etc, we do have the the console available and, of course, for those those windows VMS out there that could that can be a useful thing as well again. Here we see you know metrics specific, to how many resources that particular VM is consuming class got some more stuff over here.

B

I think the specific one that is of Interest mostly here is you know what node is this thing running on course, this is open shift. It's kubernetes everything is representable as an object. You know using custom resource definitions um within within kubernetes so and that's what unlocks all the automation um and other capabilities that that we expect from the platform um uh we do have the ability, as you would expect, in a virtualization environment, to Define uh various virtual networks that we uh are going to attach the virtual machine to so.

B

In this case, this virtual machine is attached to two networks. It's attached to uh both the Pod Network, which is um you know, the normal pod Network that that other pods within the environment are connected to. It's also got a layer, two connection, um and so that's uh that's a layer, two connection using a virtual switch on the the worker node that it's that it's running on.

B

So basically what I've done is I've I've set up a virtual switch, a Linux bridge on uh on each of my worker nodes um and then I defined What's called the network. Attachment definition.

B

um Look at that under here. So there's a network of attachment definition, and these things are really really simple.

B

You to create a network attachment definition, you literally give it a name. You tell it what kind of network you're going to use um in this case we can look at the Linux bridge and then it's just going to ask you hey. What's the name of the bridge interface, you want me to connect to, and um you can optionally add in a VLAN a VLAN number.

B

If you want to do that, if you, if you don't put a VLAN in then it just acts, you know as kind of a pass-through or a trunk, and anything that comes into the physical net is going to be available via that Network attachment definition. Yes,.

C

So- and this is one of those areas where the the open shift host management Paradigm takes over because if we think of something like rev, you know rev has this GUI where you can go in and sort of, drag and drop. I want to take these physical adapters and bond them together and then I want to create. You know: VLAN, uh Bridges or Bridges, with VLAN tags. On top of that, and then I assign my VMS to that with openshift like it's, not it's not open to virtualization, it's openshift with openshift.

C

When you want to manage those underlying network adapters, you can do it at install time. You know using.

C

um So if you're you can use the live ISO you can boot to the live, ISO configure it using an mcli and m2e. You can, if you're doing IPI bare metal IPI, you can put NM states which is yaml I, think you have it up on screen right now, so you can put like NM State directly in the install config and depending on which interfaces you can also do it day, two, which is the nncp you're. Looking at now, right.

B

Exactly this was done day two, and the only thing you can't do take day. Two is mess around with the physical interface that your pod network is is running on uh for reasons that are probably relatively obvious. This.

C

Is I won't, say I've ever I've, never broken management interfaces on on. uh You know other virtualization hosts that that never happens. Yeah.

B

Exactly Well everybody's got everybody's got to learn that lesson the hard way right um so yeah. This is this is what we call a node network configuration policy which you know really rolls off the tongue, um but this is this is all it took for me to create that bridge on all of my worker nodes um and attach it to the physical interface. So you know Define the physical interface that there that it's attached to I've, given the bridge its name, yeah I, told it no don't run DHCP on this.

B

um You basically don't use ipv4.

B

A

B

Really really quite uh straightforward: I mean the only thing that people are coming from Legacy type. Virtualization environments are, you know, again coming from, say A rad they're that they're not like that GUI isn't there in the console, because we don't really do Post Management in the hospital.

C

Yeah and one thing I noticed there is you: don't have vlans to find there because the vlans it you I, guess you you can Define them there, but somebody in the NM State Engineering team once told me that they suggest that you create vlans with the network attachment definition, not with Linux Bridges themselves.

B

If nothing else, if you were, if you were doing the vlans at NM state and in NM State, just to be clear, the NM stands for network manager yeah, um so NM state is literally configuring network manager. So you can do basically everything through NM state that network manager can do, and you can't do things that network manager can't do um so.

B

If you wanted to create an interface on a hose that was VLAN tagged, that's NM state, if you, but if you want to trunk a bunch of vlans into a host and then allow the at a at a higher level, allow sort of the virtualization users to say: hey I, want to connect this VM to this VLAN. You know then yeah you would do it through the network attachment issues. In that case.

C

B

C

One interesting thing is that the network attachment definitions along with the VMS themselves are namespaced, so you can control who has access to what in that way as well right.

B

Network attachment definitions, I believe, can be created at a cluster-wide level as well, so you've got the you've got the option um if I'm not mistaken, okay, um but yes, they are, whereas obviously anything done on host configuration.

B

The host configuration yeah all right, so back to RBM network interfaces, you can attach disks detach disks Etc. We can take snapshots of the disks, so you can see, we've got I know what you'd expect from a from a virtualization console in terms of you know, managing creating and managing a virtual machine.

B

um The one thing that is frequently asked about is whether we have live migration capability in openshift virtualization, um which is for anyone who isn't familiar with it. It's it's basically what we call our version of what of the emotion in this case.

B

um The answer is yes, we do and in fact Red Hat has, to my knowledge, never shipped a virtualization product that did not have that capability. That's uh you know. We recognize that that is kind of table sticks if you will to to do virtualization, um but it's there's act. It's kind of interesting we've got this vert launcher pod right, which we just looked at, and we know that in kubernetes you can't live migrate upon. So let's look at how this this works. This is one of kind of the tool hacks.

B

If you will um that we've done so I'm.

C

Glad you're looking at this, because this is like I love the way that we solve this problem, because once you see how it works, it just makes sense.

B

B

Okay, so what we see here is I my vert launcher. Pod is running right now on ocp4 worker 2.. So, let's live monitor. We can initiate that from the console.

B

C

Different node.

B

And what we see here is we are creating a new launcher pod for that VM over on a different host, and so what we do is we spin up an Uber launcher pod, it's running libert d and then the operator tells those two liver D instances to migrate, the virtual machine from one to the other um and eventually that should complete.

B

And then this existing one on the previous node should go into a completed state.

C

Yeah yeah, it's it's just it it's beautiful in its Simplicity right, because it's literally I'm going to schedule a second pod and then I'm going to use the native qmu libverts mechanism to live migrate. The VM from within those two things like we didn't have to do anything special or extra inside of openshift or kubernetes. In order to you know, you can't live migrate, a container you can't live migrate, a pod, but we can schedule a second one and we can migrate whatever's going on inside of there to a different host. It's just.

B

C

B

So the answer is the the live. Migration slide that I want to create is just one one sentence: big red text. Yes, we do it.

C

Yeah um and again apologies for butchering names, um so no openshift virtualization only runs on bare metal nodes. Can those be added to a pre-existing IPI vsphere cluster.

B

And the the answer to that is: yes, um if you've got a bare metal cluster and you go into the operator, Hub search for virtualization, the openshift virtualization operator would show up there. It's it's simply. It's part of openshift virtualization. It's not well.

C

B

C

You wouldn't be able to add bare metal nodes to an IPI cluster.

B

Though, oh well, if they're, okay, I was assuming a bare metal, IPI cluster and you wanted to add.

C

Yeah virtualization I think in Fahad. Please correct me: if I'm wrong I think it's, they deployed a cluster to vsphere using ipis. So it's an Easter API cluster. Can you then add bare metal nodes to that vsphere API cluster? That.

B

Is definitely a different question. Yeah.

C

That one is no yeah, you can't mix Cloud providers or a cloud provider and no cloud provider. So the vsphere cloud provider requires the node to be well a VMware virtual machine and bare metal would not qualify as that and that's actually that's an upstream. That's a kubernetes limitation I've got the link for it here somewhere um I've had this book I've had this bookmark for like three years now,.

B

Hopefully that answered your question, regardless of which the question which question you it was: okay, all right, so we've got that now there is. There is a subtlety to all of this wonderful stuff that I wanted to uh I wanted to point out, and that is to do with the Pod networking.

B

So let me go back. Let's see.

C

um While you're typing uh Sammy I'm, not sure I understand your question uh is micro, if is migration supported between just between two nodes, um so a virtual machine can only run in one place at a time, so you couldn't migrate or like create go from one instance to two instances running somewhere. So.

B

You also can only live migrate between nodes within the same cluster. Sometimes we get asked whether we have some sort of cross cluster live migration and just like every other virtualization provider in the universe, we we do not.

A

Yeah I think that was really more of the question. I I think it was or I I think that Cindy was asking like yeah. Okay, there you go yeah.

C

Between two clusters, yeah, no.

A

C

One thing: while we're on the subject, one thing I thought was interesting: is the live migration? It's just a yaml object. You can, you can trigger a live migration from the CLI just by creating a yaml object.

B

C

Name, it's like you know, virtual machine live migration or something is the name of it.

B

Yeah and in fact yeah I mean I like to um it's a console for for demo stuff, but it's openshift. Everything that we do in the console can be done via the API, um just by definition.

B

um So that's an important point. So here we're back in our vert launcher pod now and I wanted to look at the networking. That's yeah I have that, oh, but how about this config in here? No all right!

B

So let me find what you'll see here now we, the important part, is: we know the VM is attached to the pop Network. The Pod is also attached to the Pod Network. The VM is running in the pod.

B

Has a single IP. If we go into the console, go look at the Pod it'll show us the Pod IP, the Pod IP that looks like the Pod IP right there, so the Pod IP runs it lives. If you will in the vert launcher pod, it does not live in the VA. So let me go into my fancy. Schmancy consult here for this virtual machine.

B

Let's look at the network configuration as seen from the from inside the virtual machine.

A

B

So you can see here that it's got a cup. Well, it's got a couple of them, so the eth one that we see there with that 192 168 123., that's the layer, 2 Network! That is connected too um the secondary Network. So eth0 is the the Pod Network and what you'll see in in every VM here um every VM running in openshift virtualization is that it'll get the same ethd or IP it'll get that 10022 and if I look at the default route, it's default route is through 10, 0, 2.

A

B

And it shouldn't be a shock to see that ten zero two one is there in the vert launcher pod. So what we're doing to get around the fact that we need two entities, the Pod and the VM both connected to the Pod Network through a single IP address, is we're using network address translation. You know, just just like you use at home um to attach multiple things behind a single IP address, so so do be aware of that.

B

um Let's see what else did I want to show ah I wanted to show the a quick run through of provision creating a new virtual machine I, don't think I actually spin one up, but I did want to show what this looks like select, which template you want to use generally you'll want to customize it somehow all right. Let's see this is where I like to set a password to something that I actually know so I can access the console.

B

um One thing I do love about this is that it tells you what the uh what the username is, because I always forget that different operating systems.

C

It's really nice in the console view too, for the VM, where it shows you the credentials. If you use cloud in it.

B

A

Right so and then, if you does tutoring demos, so if.

B

You go on to this. If you go on to this uh this screen. This is where you have the opportunity to say or delete additional network interfaces on different networks. I should note that, in addition to the bridge or v-switch based layer, 2 access, we do also support sriod with this. So for, if you've got those those very high performance um workloads, then you can use SRI OV with that. Just now you you do not have to attach a VM to the Pod Network.

B

um We certainly think you're going to want to in most cases, um but you're not actually required to do so. You can. You can delete it here. Virtual machine will start up just fine.

B

All right, let's cancel that yeah, Ah that's not what I wanted to do all right anyway. You also have the opportunity from that screen to uh to add additional disks and stuff, so do doing the basic customization that you'd expect change the number of CPUs change them out of memory Etc. um One nice thing: you see this this Source available flag here for operating systems, for which a a uh you know we can. We can do this.

B

um We automatically download the the images for all of those operating systems, so you don't have to go out and and manually download them. um You can also turn that off. If you don't want your users to be using the you know, stock ones from Red Hat and restrict what they want to have access to.

A

I was just going to say if so, if you're doing this in a disconnected environment, is that like, when you're pulling down the operator? Is that part of that operator pull down? Or is this like a separate step that you would do um outside of like like an OC mirror and then just pull down openshift virtualization? You.

B

See mirror OC mirror should bring it down so okay, then it should be available to you. Okay here, if that's not the case, I, don't think it is.

B

um Let's see we also have this very long list of pre-populated templates, which is you can see, are not in my default namespace, but you can see so. You've got basically all the operating systems with various size variants of them, and of course you can. uh You can customize all of those one thing that I should probably also show yeah wrong.

B

Straight from catalog fine is.

B

In the customized, aha, okay, so you can change the source that the virtual machine pulls its initial disk image from so of course, you can use the default that is specified from in the template, but you also have, as you can see, a bunch of other options, so you can just simply clone an existing PVC.

B

um You can pull an image from an image registry.

B

um Basically, we can take like a qcal essentially and wrap it up in a in.

A

B

um In a you know, cloud cloud or container image: if you oci that's the word I'm looking for, um and you can put that in an image registry access it from there, you just pull it from any old HTTP, https URL um you can upload it.

B

I guess upload it from your local system through the browser, never tried that one don't know why you'd want to, but you can do it um so you've got a lot of options there um and what I think is is potentially quite interesting is his ability to use an image registered because again we talk about these are pods. These are kubernetes workloads and we shouldn't have to care whether a given thing is a container or a virtual machine.

B

We should be able to use the same processes and such with that, and so one of the things we should be able to do is use the same processes to update and manage the set of images that we use to provision these things. And so you know one one thing that you know you can do that: we've got people working on and is use tecton um openshift pipelines to maintain to build tests. You know update your virtual machine images, just as you might do for container images.

B

Those can then be stuck into the registry and and used from there. So again you can use exactly the same processes that you use on the on the containerized workload side that you do ever use those on the virtualization side and.

C

I want to I want to poke at that for a minute right, because so with with libbert, with with a KVM based virtual machine right, a disk is more or less just a file. I mean yes, we can pass through a block device and stuff like that, um but so a disk is more or less just a file.

C

um So when you say you're using a container image a registry to store those disks effectively, it is literally a container image right you when you want to store one, you create a new, you know container file or Docker file, and you say from scratch, and then you add in the Q cow or the raw disk file, that is, that template image to that container image.

C

You build the container image, you push it to a registry and what happens when you use that Source disk? If you, if you want a persistent virtual machine, it basically pulls that container image in and it copies the disk into a PVC and then that's how it runs. You can have an ephemeral virtual machine. You know where, when you turn it off it, it just like any other container any other pod. It removes that disk and everything is reset.

C

But that's why, like the Technologies, the the terminology, sometimes misaligns, and it makes it more confusing than it actually is,.

B

um Absolutely we yeah I didn't really go much into storage, but the way we do, storage, as as Andrew said, is that or every persistent at the end, which is which is most of them. These days um yeah. We we just do a PVC, so it's whatever your cluster is set up to use. You know, as far as storage, whatever csis you're, using to provide storage to your containers.

B

That's what kubernetes is kubernetes, that's what openshift virtualization is going to use and you can tell it what storage class you want to use um and it'll just it'll. Just do it! You, um generally, you want block rwx um for for virtual machines. um Rwx, you need the rwx in order to do the live migration, yeah.

C

And a lot of people may pause when you say wait, you can't do rwx with block with block PVCs. So actually you can't the vast majority of storage providers do support rwx with block the key. Is you have to use a block mode? Pvc, not a file system mode, PVC, so file system mode is exactly what it sounds. Like so say it's an iSCSI. You know volume all right, iSCSI PVC, using file system mode. Essentially the CSI provisioner will Mount that storage device and then format it with a file system.

C

You know EXT or xfs or butterfs or whatever you tell it to format it with uh that will still work. You can still use that with openshift virtualization, but it's a read write once with basically every storage provider that I'm aware of at that point, so block mode it effectively. It mounts the storage device, so it's underneath you know Dev whatever, and then it passes that device directly into the pod, and so you know I alluded to this before of like with qmu with libert.

C

You can just attach a device on the host into the virtual machine. That's literally what it does it just it, it doesn't care what's going on in there or it doesn't have any awareness, I should say of what's going on inside of there.

C

Qmu is doing its thing with it and that will allow it basically libverts and qmu manage so that only one host at a time during the live migration process is accessing that block device so that it doesn't have any risk of uh of corruption, and, quite literally, this is exactly how it works with Rel. If you've got two Rel hosts and you're running, you know you're using virtual machine manager. If you want to go way back to the basics, and you want to take a VM and go from one to the other.

C

You just Mount that device on both sides and tell it live migrate and it'll move the state of that VM from one to the other, and it works exactly the same. So I always thought that was super. Interesting of like all. Many of these principles are the same thing. You know if you're familiar with Rel, virtualization or rev or openstack like it's just the same thing. It's just abstracted using kubernetes mechanisms, um I I every time, I dig in every time. I learn something new about it.

C

I'm always surprised at how straightforward it is once you get down to it.

B

Yeah, it's that's the funny thing it it's like Simplicity, wrapped in complexity or something I, don't I, don't know or or something like that. So.

A

B

Are we're getting we're getting short on time uh here, so I and I did want to show this um in the same in the same game, and this goes back to the the question that Manish asked very early on um about replacing VMware or or another traditional virtualization environment, with with openshift virtualization um and as I kind of said like yes, it can do it it for 90 something percent of virtual machines in the world.

B

um We can run if it can rhyme in KVM. We can run it in openshift virtualization.

B

um That said, this is not a direct replacement, openshift virtualization and its Upstream covert project from from which it's derived you know, are all about bringing virtual machines as first-class citizens into into kubernetes or into openshift. They are not about making kubernetes are open shift work like a traditional virtualization environment.

B

um So you get the power that you get with kubernetes. You get get Ops as I've I've shown here, where I've deployed a couple of virtual machines based on the contents of of a git repository that defines them um you can you can use techcon to build your images. We didn't really show it, but you can once you've got your virtual machines in there connected to the Pod Network. You can expose them as Services as as node ports.

B

If you want through load balancers through routes, Etc, you can use service mesh to you know, do blue green testing to do Network tracing. So you can do all this. Stop all this cool stuff um that you can do in the open shift kubernetes space, um but it is a transition. It's not the same as using those those older Technologies.

A

And, what's a what's a use case where- and this isn't meant to put you on the spot but like what's a use case where um open to virtualization doesn't fit like what what VM use case wouldn't fit on this? So.

B

So there's you know: well we don't we don't do false talks. Vmware has a a fault. Tolerance feature that literally replicates in real time the state of a virtual machine from one host to another, um so that if one of those hosts dies, it's you know it's it's basically seamless or, as close as you can get to seamless.

B

um That has a lot of performance implications. Obviously, you'd only do that for really really super critical stuff I. My understanding has always been that that's not widely used, um but that, as far as I know, is a unique feature to them. We don't have it in in any no.

A

B

Else has that, as far as I know, um there's support considerations um to my knowledge. You know you got like your saps and such out there that get very, very Twitchy about what they will support. So you know, as we've talked about out and shown these virtual machines and openshift virtualization are using the exact same kernel, KVM qemu libbert.

B

As you know, we use on Rel that we use in openstack that we use in Red Hat virtualization, you know, but an isv can still decide that they don't want to support it because there, because it's Tuesday or something like that.

C

B

But it's it's much more! It's really much more about it's! Not it's very rarely about! What's running in the VM, it's about what how you Mr administrator, want to manage your Fleet of VMS. You know: do you want to continue to use your lovingly crafted ansible playbooks? You know that you've you've written over the years to to deploy and update and stuff things. Or do you want to use the the same Technologies and and operations that you're using in in the openshift kubernetes space to manage your more modern workloads.

A

Gotcha, thank you. Yeah I I think that this is awesome, I'm glad you showed the Argo aspect of this because I think that, like as image, Builder gets more, um you know more notoriety. People are using it more and more I think things like openshift pipelines and Tech time and then, like that whole you know CI CD pipeline to now, I can build my images and I can deploy them and test them all in cluster and then I can you know I can go through a promotion process using git, Ops and I.

A

I think that it's this is awesome, I'm, really glad that you showed this. So thank you. Yeah.

C

Very well and Ian I know you have to drop here in just a minute, um so Johnny and I will we'll stay on we'll we'll keep uh answering questions and chatting with folks, so we'll be here for at least a few more minutes, um but I wanted to take the opportunity, um as the clock takes over. uh Thank you. Ian really appreciate you uh you're coming on showing all this to us.

C

It's been great to get an in-depth look at this um like I, said I, I love how it feels so foreign and complex on top. But when you look underneath the covers, it's just KVM, it's the.

A

C

We've been doing for the last 15 years.

B

C

B

Just KVM at the bottom and it's just openshift and pods on the top yeah.

C

B

There's a bit of magic in between.

C

B

Right I do have to drop thanks, guys really enjoyed it. Thank you. Thank you.

B

C

You all right uh all right there we go.

A

That's awesome good.

C

To see you full size again, Johnny yeah, I, know, I, know you too yeah it's um it's funny right, because I've been doing virtualization for I, don't know, I, think 2007. It was the first time I I started with virtualization. Like really you know: meaningful professional career virtualization back in the days of ESX 3.0 um and like I I've I've used vsphere I've used rev I've used openstack.

C

um You know, I've I've experimented with I've done quite a bit of of toying with openshift virtualization and getting familiar with it and understanding how it works. And it's really this.

C

It is a change in the control plane in the same way that you know rev versus or virtual machine manager on Rel versus or now cockpit, I guess um versus um uh you know: rev and the Rev manager versus openstack and Horizon. If I remembering the project correctly,.

A

Yeah for the UI.

C

They all provide this. This sort of abstraction and control, plane, type of thing of you know in rev, I go in rev manager and I click, the button that says new virtual machine and it pops up this form. And it's what do you want to name the VM and what you know do you want to create a disk? And what network do you want to connect it to?

C

And all these other things and at the end, when I click create it more or less, creates an entry in a database right and when I say start that VM it uses whatever logic it has and says: I choose host three host.

C

Three, you know here is a XML lib, verx virtual machine definition go start it and you know, essentially it's the disk is in a data store, that's already attached and the network that it's going to connect to is a network that already exists and all of those things openshift virtualization is more or less the same.

C

When you create that virtual machine, it's not creating an entry in a database, it's creating a yaml definition that ultimately gets stored in SCD, which SCD folks will be quick to remind me: it's not a database, it's a key Value Store um and when you say start this virtual machine, it's scheduling a pod right. It it like, like Ian, showed it's the vert launcher pod. It says start this pod and mount this.

C

You know Mount this PVC or or these PVCs, because that's where the disk is, you know, connect it to this multis network, because that's how it's going to connect to the world or the openshift sdn.

C

um uh So to me, it's it's really interesting that fundamentally they're the same just the implementation is, you know, it looks, feels taste smells different.

A

um Yeah well, it's kind of like the whole thing, with kubernetes in general, right, like the way that we handle applications, the way that we like access them, the way that we deploy them and manage them. It's the same thing now with just now virtual machines right like where, before we'd, give it an IP and we'd, go directly to that IP. Where now we use a route and uh yeah, like you said, like we use a pod to configure that front end for it and yeah it's it's awesome. It really is.

C

Yeah when I talk to people you know or when I talk to you know anybody whether it's uh you know somebody internal somebody external about openshot fertilization. The question always gets asked, you know: oh, can I replace you know vsphere or rev or openstack, or any of them really, and the question is always well.

C

Maybe it entirely depends on what you're doing with those VMS, and you know it's if, if you felt comfortable going from say rev to openstack or openstack to vsphere, or something like that, you know it's a it's a paradigm shift in how you do virtualization management, then in theory you could replace it. There are other things that are different about it right for a long time. Openstack didn't support, live migration.

C

um You know I feel like a long time is a little unfair now because it's probably been longer now that it's supported it than it didn't support it before. But there was like years like four or five years at the beginning of openstack's life cycle where it didn't support live migration, and so you know there are different capabilities, um as as Ian pointed out, like fault, tolerance, isn't a thing. um You know, I think the way that virtual machine restarts work or ha is within the constraints within the Paradigm of kubernetes.

C

um You know so we have to and I think we've talked about this on the stream before poison pill operator yeah. You know you have to have something that effectively fences the host and ensures that the disk is disconnected because you don't want it to be mounted. You know. Oh, it's just a network partition. It's still there.

C

It's still running it's just the management plan can't talk to each other, so I don't want to start a new VM, that's using the same disk because now corrupted so again, rev and openstack have had these concepts for years.

C

um You know you still have to have those mechanisms in place, but it's just KVM I, don't know, I, always thought that was really interesting um again, just the the Simplicity of it. Once you get just scratch, you know, lift the covers up one level yeah.

A

um Yeah exactly you just look: look underneath it and you're like okay, I'm comfortable with this I I know what these are. um I I think, like the I'm glad I showed. It too was like the way that the the um the launcher pod manages the sdn IP, and then you get into the actual virtual machine, and you can see your you can see your KVM Network and then you can see like the the um the address. That's uh communicating back into that pod back into the cluster.

A

You know, I mean I, think that that's unique um and that yeah definitely like if you're trying to like come into it cold. It's probably like hell in the heck. Is this happening? What is going on, but not like seeing that, like that workflow, it makes a lot of sense now. You know, and um I I did want to chime in on something that our hope9 said uh about video. So video is like a it's like impression: uh yeah.

C

I was looking at that acronym, like that, seems so familiar and I I kept wanting to put like you know, VMware data, something, but no it's the it's the Rel. It was a Rel. It was an acquisition that red hat did yeah and it was layered on top of usually lvm, and it would do things like um you know: data compression data deduplication, as data was being written.

C

um I I used it in my lab for a while. Now that I think about it, um yeah and I had the I was a glutton for punishment. I was using like manually configuring, the iSCSI, Target and issue.

A

C

Presents you know luns, you know, which are basically files hosted on a video object out to my my lab host yeah, that was, that was fun. Yeah.

A

I I I'm only familiar with it, because I used it with openstack back like four years ago, I think so three three or four yeah four years ago. So I don't know at some point in the past and um I I, don't know if the um volume group operator, I I, think kbm supports it. There's no doubt about that, but I think that how you'd enter how you'd get it to work within the cluster.

A

You know like because with the volume group operator uh or the local storage operator, you know, or the odf local storage I'm, just not sure um how you could get it to work. Maybe you can I'm not sure I.

C

Don't know it's a good question, I hope, I, so I don't think you can use it directly um because there's no CSI provisioner right. So there's no CSI provisioner that you can point to. You know: here's a Rel host, that's running video that I can carve up and use for icecuzzy, lens or or NFS exports. There's no CSI provisioner! That works with that that I know of um if you were to run it on the Local Host.

C

So let's assume that I had you know: I could get lvm and video running on like a single node openshift instance. You could use the local storage operator to carve that up and use Virtual Machine disks then, but that's a big. If I don't know if video would run in core OS I doubt.

A

Yeah yeah, that's where I was like I'm, not 100 sure but yeah. Definitely it supports. Kbm supports it. Yeah! No doubt about that. Yeah.

C

um Chandler not sure if Ian mentioned it, but uh big open shift win is that VMS can use services and connect with them yeah um so that that's one of those things where so a VM, an openshift virtualization virtual machine can connect to the openshift sdn or it can connect to you know directly to an outside L2 Network right, a VLAN Network, just like any other VM that we've used for the last.

C

How long has there been virtualization 30 years now um you know so it can work either way um or it can do both right. I can have a VM that has one interface, that's on the sdn and 10 interfaces that are connected to different vlans if I so choose. So for that interface, that's on the sdn. Remember it's just a pod! So I can create a service that points to that virtual machine and then other pods can access it. Using that abstraction I can create a route to expose an application. That's running in that VM.

C

Just exactly the same um to me, that's a huge win for migration or uh modernization stuff. So you know: hey I'm running my I. Don't know three-tier web application. You know, and today I'm going to take um I feel like I'm dating myself here, like you, know, I'm running a capacity in a virtual machine and I'm going to import that virtual machine directly into openshift. Today, I'm going to put a service in front of it and it can do its thing tomorrow.

C

I go and I create a containerized version of that I can deploy them side by side and I can use my service or even the service mesh, because virtual machines can also connect to the service. Mesh and I can direct traffic to both of those and then I can say. Okay, my pod, my containerized version. It's working great. It's awesome, I'm, just going to turn off the VM and again the service mechanism handles that, excuse me and and is able to take care of it.

C

So I thought that was really awesome um to tablet to Tavernier. Again, apologies for butchering names. Can you suggest a DIY Home, Server builds getting into pen testing and whatnot need a roadmap for what to get um that is going to be like super highly opinionated, um because it really depends on what you want to do so.

C

Pin testing um you know could be. Multiple operating systems could be network devices could be various appliances and all of that other stuff.

C

So and it depends on what those are, how complex you're set up and you want it to be, and all that other stuff um somebody in so their Red Hat has a internal group. That's really active. The home lab group is really active and yes, it is somebody. There pointed me to a website called lab gopher over here in twitch, um so lab gopher is basically an interface to eBay um and it you can use it to kind of search for and find different servers.

C

You know obviously secondhand it's eBay um with different characteristics, so it's a good way to find hardware that um you know potentially at a discount all the normal caveats. Etc of you know, purchasing secondhand servers apply, um but yeah it's you do have to be a little bit careful because, as we, you know in the later versions, now of rel8 and of course, rel9 going forward, we've started to remove compatibility with KVM for older processors and by older processors.

C

I mean things that are like uh 2009, 2010 and earlier, like first generation opterons, and you know all that stuff. So you know keep an eye. You know. If you really want to be. um If you want to guarantee compatibility, you would want to look for something. That's on like the rel8 hardware, compatibility list or well-9 Hardware compatibility list.

C

um Yeah ask Noah. Also has some level one tech center. They do a lot of stuff with home lab and all that other stuff. um So yeah there's a number of other other folks that can help with that. But personally.

A

Yeah I'm sorry I just saw the name so I was like yeah.

C

Yeah so me personally, I use um old desktops, uh so I I, one of my hobbies, is building building computers.

C

um So, like you know the gaming computer that I use and stuff like that, so I just have old versions that I repurpose into uh into home, lab machines and just bump up the RAM and that type of stuff. So it just depends on what you're familiar with what you're comfortable with other folks, um Christian I think Christian went with like a R7 40 or something that he has running in his house.

C

So the the downside, um at least for me, so my wife, who um is is more sensitive to noise than I, am I've spent a lot of time in data centers, I'm, partially deaf, so I, don't I, don't hear the the fan noise quite as bad, but my wife is very sensitive to the fan noise of servers. So you might. You might also want to take that into account if your family is not uh is is sensitive to that, as well as the cost of power.

A

Yeah yeah I've got a I've, got a T440 and then I've got some super micro, E300 8ds they're, like the little pizza box servers and um yeah they're they're awesome, but they are so loud. Man like the case on they're loud, like you can hear it downstairs. My wife is always yelling at me about it. So what's what this guy Sean Nelson, who who's chiming in here? If you were to see behind him in his in his uh apartment, his closet is full of like Hardware like this dude's got all kinds of crazy stuff.

A

This dude's insane.

C

Yeah yeah I I definitely don't want to um understate the cost of power, though um you know, and you're in Texas Texas has a weird power thing. You know and yeah how much it costs and all those like power can be, and it's not just like the direct power to run the servers, but also remember it. It gets turned into heat.

C

um You know we're literally taking two rocks smashing them together and sending electricity through it. It generates a lot of heat, um so you have to do something with that oftentimes. That means air conditioning. Well, now that's an additional cost, and you know all that other stuff so yeah, which is what I was watching um some YouTube channel the other day, and they were talking about how, like the heat, wasn't working in in the house.

C

So instead he just stopped venting the server rack outside and was venting it into the house and it kept it like just warm enough to be thoughtful. Like wow.

A

That's awesome, yeah.

C

Whatever works, I guess right, yep.

A

So one last thing on that. uh The other thing to keep in mind too, is if you're wanting to do like VMware VMware is very restrictive on like their Hardware list. So and that's not like a slam on VMware or anything like that, but like they, they generally don't um support like I, don't know if it's like two or more older models.

A

You know so like the 440 is the fourth gen, so they might be on like if you're on, like the 460 or something like that, you know mine might be out of a out of support. But just keep that in mind. If you're wanting to do VMware that like check their compatibility list before you buy the server because that'll be some heartburn, if you buy it and you can't use it, yeah.

C

Yeah um so I'm trying to scroll back up through here, see if I see any other questions or anything that we might have missed um so just before we uh close the stream. If anybody has any other questions, if we missed anything, um if there's anything that comes to mind, please don't hesitate to ask those questions now. I'll also say um and I'm realizing as I'm saying this.

C

That I didn't do this earlier in the show for anybody who's not watching us live I know we get quite a few folks who don't watch us live, uh feel free to reach out. So you can reach out to me at andrew.sullivan.redhat.com I love to get email from folks. We get email on a regular basis. So you know anytime, reach outs. Ask us questions. You know, ask for clarification, ask for resources all that other stuff, we're happy to to help us. However, we can. You can also reach me on social media.

C

I'll admit that I have not been on Twitter very much lately for reasons, but I am there and it does eventually send me. Emails and I try to keep up with those, so at practical Andrew, but Johnny and I are also on uh the kubernetes slack. So there is a openshift users channel on the kubernetes slack, uh we're also on Reddit.

C

So if you go to the openshift subreddits, we both frequent there as well so email, andrew.sullivan, redhat.com or johnnyjo-n-ny redhat.com, and whichever one of us, if, if you don't remember both, you can always just send it to one or the other, and we always add each other on. If, if one of us is missing.

A

So we did have one question and it was from our hope9 and it was about having a VM running in a service, mesh and I. Believe the answer is yes, and if Chandler is still listening, uh Chandler I, don't know if you know for sure I know, like I think when you were on a couple months ago, you were talking about at least building that out, where you're building out the service mesh with the VM on top of it. So I think the answer is yes, it does work. Yep.

C

Yeah yeah, it does work, um they added that in 4, 10, I. Think, okay, if I remember correctly, um you know I'm I'm, a virtualization nerd, so I've been keeping up with it for a minute. That's awesome, yeah!

C

So all right um yeah! So again, if anybody has any uh questions, anything that comes to mind, don't hesitate to reach out. Thank you so much for joining us today. We really do appreciate it and uh joining us after we've been on vacation. If you will, after we've been on Hiatus for a couple of weeks for the holiday and then uh what was it we had. There was a conf like of the last five weeks, I or six weeks, including this week. I think we've.

C

This is our second stream because it was like overlap with ansible Fest and then overlap with the what's next for openshift, 4.12 and then the holiday, and then our botch stream. You know our botch topic last week and so we'll be back next week, though um definitely be back next week, but that next week will be the last stream of the year and then we'll be back in January.

C

For, after all, the holidays and everything um yeah I know, I got a nice long road trip to kick off the holiday for me, going to a wedding down in Florida.

A

Oh dang, nice so.

C

Yeah not very often I get all dressed up. um You know I I I've been in covid. You know status for the last three years. Like you know, it's it's like gym clothes every day, right, yeah,.

A

I had to put on a button-up shirt uh back in August and uh I was like I still got it put it a lot of buttons up and I was like. Oh, my God like the neck doesn't even touch. I was like this is I'm in bad shape. So my tie was all the way up. Yeah it was. It was bad. What's.

C

That movie uh Tommy Tommy Boy, you know where he's wearing the wearing the jacket and singing the song yeah yeah.

A

C

Feel you yeah. Yes, that was me.

C

All right, Jill, thank you again, um we'll see you next week, uh Johnny any last words.

A

Real quick I'm getting the link from Chandler for the service mesh and openshift virtualization blog. So if you're interested check that out, um we had Chandler on a little while ago, uh yeah he did some awesome stuff with openshift virtualization and uh get Ops. So I think he's on your team Andrew. um You know. So it was cool to have him or at least no.

C

Chance Chandler's uh Chandler's in engineering he's on the integration team. So oh.

A

Okay, okay, they do.

C

They do lots of cool stuff where it's um like. The the so engineering says. Oh yeah, we're going to add compatibility for service mesh Chandler's team are the ones who actually go out and like prove that that works and then show the rest of the world how it happens. So.

A

Nice: okay, that's what that's what it was I remember now, but yeah! That's it man! Thank you! It's good to see you again, I'm glad I'm! You know I'm glad the IBM thing is done and uh yeah to Jamie's point like I, don't have to host the show by myself anymore and they don't have to put up anything so yeah, I'm glad you're back because uh we're a dynamic duo, not a dynamic single, so yeah.

C

All right have a great week, everyone, foreign.