Cloud Native Computing Foundation CNCF Webinars, 30 May 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Webinar: Meet KubeOne: An Open Source Cluster Lifecycle and Operations Tool for HA Kubernetes

Description

Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io

Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects

In search for a feature-complete solution that supports HA clusters, follows Kubernetes best-practices, and comes with a simple and declarative API based on the Kubernetes Cluster API, we could not find a single one that fulfilled our needs. Therefore, we decided to build KubeOne.

KubeOne takes care of installing, configuring, upgrading and maintaining Highly-Available Kubernetes clusters. It works out-of-the-box on any cloud provider, as well as on-prem and in bare-metal environments.

In this webinar, we’ll introduce KubeOne and show you how to get started.

A

All right we're going to go ahead and get started now. I'd like to thank everyone who is joining us today. Welcome to today's cmcs webinar meet cube one, an open source, cluster lifecycle and operations, tool for high availability, kubernetes, I'm, Kaitlin, Barnard marketing manager at CNCs and I'm going to be here to help moderate.

A

Today's webinar I'd also like to welcome our presenters, are Tom and Marco from loofa, so just a few housekeeping items before we get started during the webinar, you are not able to speak as an attendee, so there's a Q&A box at the bottom of your screen. If you have any questions throughout the webinar, please feel free to drop them in there and then we'll get to as many as we can. At the end, the session is also being recorded and will be sent out afterwards, along with the link to the presentation.

A

So with that I'll hand it over to Marco to kick off today's presentation.

B

Humo everyone thank you for joining us. Today. We are going to talk about q1, a new open source tool for managing humanity's custard. I am mark modern age I'm, a software developer at Lucy at computer science student at the university of belgrade. I am one of the core maintenance of the cuban project as well as cuban. It is contribute or contributed to Castle API project and maintain of us API provided for digital ocean today with me, is Artyom.

C

Hello: everyone thanks for joining I'm, Marcin, also major contributor to q1 and welcome Marco.

B

You can also find us on Twitter, github or kubernetes. Like me as x4e, and our team is kron4 eg.

B

We are going to talk about what is coupon, how it works and to do a quick introduction will see some core concepts and how we did did some stuff, and then we are going to have some live demos to see how you can get started with the custard.

B

We can manage human, a discursive on Amazon and we're going to explore some more advanced at Cuba, so butts q1 q1 is a tool for managing humanities class, the lifecycle, but across the lifecycle we mean installing it promise unique humanities upgrading to the next minor or batteries and our provisioning the costs, which means that we destroyed the marquee nodes and revert what is done with you. One q1 is open source tool. You can find it on github.

B

It is very now sure it works on the most popular cloud providers on on premises, it'll bare metal he supposed cuban. It is hydro, variable, casters version, 1.2 rating and offer.

B

Among many tools out there, we are often asked why we build coupon invite q1 is so different than any other tool. Human. It has brought us a new way to manage our workload. We are doing it in a different way than before, like decoratively, but manager. Kubernetes clusters is still a hard task to some degree. There is still a lot of manual steps involved, a lot of scripts that are not really cloud native and we wanted to change that. We wanted to apply resource, run it managing workload to clusters.

B

Beside that, we needed a solution to quickly get started with the cuban I discussed because for our management solution, cuba, Matic the it runs on cuban, it is itself and in surge for surf solution. That is also feature complete. We decided to build cuba.

B

Cuban uses many AA test technologies to bring many features in an easy to consume manner. Such technologies include cube a DM cube romantic machine controller, which is also open source. If you use user to manage to party nodes, cluster api and many other stuff, given bricks, decorating class to a visitation, all clusters are defined in a form of configuration manifest. You can share such manifest in version control systems with your colleagues but most important again. You're such manifest to easily reproduce the costume you get before many times you want.

B

You want, provides ready to use clusters. That means that when we proviso cluster reconfigure features are selected, we deploy CNI plugin, we create more key nodes and give you cluster that you can just use optionally with q1. You can configure various features for security policy, dynamic out the clock, back-end metric server and more when we build cueball, we wanted to integrate with provisioning tools that we introduced before, such as tariffs or delete q1. You can do it out of box.

B

It is possible to integrate q1 with terraform and we are going to see about that later, but also you can use cuban with other tools such as ansible, Confirmation and more.

B

Then we build q1, we wanted to make a tool that works on any provider. That includes our premises.

A

B

And with you one, you can do it but to enjoy all.

A

Features like many chick.

B

World keynotes using human rating machine controller or if specific book provider has some provide a specific configuration in it or you want to automatically deploy providing specific features like external controller managers and to be able to use our outer box tariffs or mitigation. The provider needs to be officially supported.

B

Officially supported, provide this for now recruit abs, Google, Cloud, digitalocean, Kettner packet, open spec in the as of today, VMware vSphere. Alright, now we are working on support for Microsoft Azure and we hope that it will end in the next series.

B

Now that we seen what is q1, let's see how it works and how we decided to build it.

B

D

Uses many tools, Solutions.

B

Is building blocks mostly important, Jubei TM is used to provision, enjoy cotterpin notes a Hadoop cluster updates, but we use government or shrinko dollar, which is basically cluster API, to manage working notes.

B

Then you run some of human operations like install upgrade, or he said the environment is repaired over SSH. That includes installing and upgrading binaries configuring needed components running cube ATM, but as we have the cluster, we are using. The client go cuban. It is library to deploy, borrow use manifest such as CNI metric server, machine code or add more.

B

Probably the most important process is the install process and we are going to start from there to learn how it works and how it looks like.

B

The user is expected to provide infrastructure themselves that could be done manually or using any tool such as terraforming, simple and more in case of terraform, as mention that we have integration which reads the telephone states and pickups all needed information and provides that to q1. We are going to see later how it works and what information you need to provide, but beside infrastructure, you need to provide the Kuban configuration.

B

The configuration manifest put this information about the cluster in predict, humanity's version. What features you want to be enabled on what cloud provider are you running and more also, here these two key pieces. We are going to provide that to cuban and there are four main steps.

A

B

Happens when you run through one instant, the first is provisioning. The cluster that accrues history. We needed binaries for Cuba natives, as well as installing the doctor Oh see here. Binaries we generate you. Pdi file that is needed to join code operates nodes and we need them and began certificates needed for hugging available clusters. With configuration file for cube, idea and certificates in place, we need to size them between all control, paid, mods and office. I classic is done. We use cube, ATM twin it and join photo playing notes on stat is done.

B

We go to next step and that is double cube. Config file queue, coffee keys, later use it by cube war to deploy manifest, but also it is saved so operator can use the cause rate, and then we go or coffee go and deploy step very deploy.

B

Cni configure all features you requested, deploy cloud, provide their specific features like a signal called couturier manager, deployed machine controller and then create working notes, said Michiko Toller, who manages working out he's basically Castle API ed, but with caster API, were, can also manage it by creating machine deployment, object which works just like basic appointment, but instead of managing workload and pods, we manage machines and machines of the cloud provider or in Europe are also here. Machine behave faster that is ready to use now.

B

I hope this sounds nice, but let's see how it works. In practice, we are going to create the cluster or abs and to.

A

B

Have some steps? The first one is to create instances a needed infrastructure to be used by humanities. You won't comes with example. Tower for scripts that you can use to get started is a reference to build your own escapes or not to see how everything works, and for this demo we are going to use styrofoam to get started after that.

D

B

Are going to build a basic chemical figuration file that is going to get us started later on? We are going to see how more advice of perfect wage the fire looks like it bought features. You can enable.

D

B

Run a Cuban, install command and then enjoy with the new cluster, so happy cluster manifest looks like at the top of the file. You have basic information like in any other humanities manifest. This is the API version. For now this is V 1, alpha, 1 and kind. The defines that this is a open cluster.

B

Then these versions.

D

B

Verify with cuban at this version is one point for two points tool and we define that we are.

D

Running on Amazon.

B

Now, let's switch turn.

D

B

Us see how this works, so the first one is. We are going to lose styrofoam to create the infrastructure for this demo. We already have prepared this assist and we are going to transfer either from apply to see. Is everything in place.

B

Now it out put it some information and we need to export it. This is done with tarp from output in JSON format. We saw finalists for it if it JSON and now, if we ready you're going to see that it has some basic information like Cuban API, which is API at point, and usually this is small balancer we have Faust's, and this is where control plane will be deployed on. We have comm, provide the name class name, private addresses public addresses and SSH information, and at the end we have information about working oats.

B

Now terraform is not many cheek, but notes. This is done by machine controller, but as to create working nodes. There are many information like Amy RA, a mobility store or ESS profile, and many other. This is from benefit infrastructure. We provide this entire form and cue-ball is able to read this and basically create machine deployment, object that is going to trigger motion control to create instances.

B

So now, let's build the configuration file, your chance, perfect, quick comment that is printed. The basic configuration file similar to one built with of these slides- and here we already have it here- you can see we have kubernetes version and so on. We are going to apply it. This command also has much more options that we are going to export late, but now, like these components, Club to install our cluster, so what it takes it takes configuration manifest in our case of gamo and optionally.

B

If you start from integration, but the tariff of state and output, we exported, in our case young JSON. Let's.

A

B

Now it is taking steps we describe world on the graph and diagram and one piece.

A

Is running this.

B

Is a good opportunity to see on what, if SL shall be deploying this, we are going to take a quick look at our tariffs or scripts. So what do we have? We can define some aiming yard if you want to use Ubuntu as well as source information about the mobility. Sounds gate SSH key, so you can later access this. This is security group, so we limit as you just need a port. This is default communities, counterpane, api, endpoint, art and beside that we create load balancer and we create fries. This is by default.

B

Now you can change that to your desired number that you want, and it is important to take a look at the output file, and this is what, when you retire from us, what you get. If you want to your star for mitigation, it is important to follow this way of exporting Eddie's variables. So cuban can read that first, we expose cuban api part, and this is usually not balanced, a red point in case such provider. We want don't have broadband sirs, it is recommended to dedicate distance for a load, balancer and use some solution in.

B

We have subscripts, but it is recommend, have a load balancer and then we have cure costs. The defines the needed information to a deceived, sshh class name and so on and I said we have workers part. This is just to be read by q1. There is nothing that tariff or headers about workers and he defines number of replicas public key. So we option we can access the working nodes operating systems and come provide a specific stuff.

B

Now this is from machine code, oil and if you want to learn more about how this looks like and what do you need to operate here, you can take broke the machine programmer repository and in the examples.

A

B

The repository you're going to find many.

D

B

That you can see- and there you need to pay attention to the clock- provide this back part where you can find what options you can expose. It go figure.

B

Now, let's see how instant is going on, oh, it is done so what happened? We had cube ATM and rubidium inserted or free cotterpin nodes. It joined them, we don't roll, it cube coffee and then we applied kernel, CNI plugin associate I. We have choice between two plugins. This is kernel which is basically calico plus final and we've net, which is more aggressively support. Some more features such as decryption. Then we install machine controller and wait for it to come up and we create vertices.

B

So if you see what we got, we got the cube file and we can export your parking variable.

B

And now, if use you cut off their balls, we will see three notes now. This is all contraband notes and it takes some time for what a note to click get created, to get provisions and to join a cluster and I believe it will be soon. But give me time we can take a look at machine deployment object.

B

It is in your system by default namespace, and we.

D

B

Have for one it is for Africa on Amazon running wood and if we take a look at machines, so many great machine deployment is, will create underlying machine set object and then machines you'll see that we have the first machine and it already has IP address. So it is created IPR the society and we are just waiting it to join the cluster.

B

What is reported here is there with machine controller and Casa API you get to manage 40 nodes with just you cut. So that means that you can do something, such as any machine appointment.

B

System, let's say so: basically, it is plus API machine diplomat object. If there is some information, such as how is it created information about infrastructure,.

B

Tanks operating system and more but also most supported versions, and there should be a number of all this code and there should be number of replicas yeah is here number of replicas. So if you increase it, it will create more work nodes, but instead of just at it you can use skewer suppressors. Let's go see, have you got it? Yes, it is here it is already ready and you can do you can go scale, machine deployment.

B

You can system, namespace, robot or replicas. Let's see three and now, if you take a look at machine deploy nice, it's a spring replicas and the machines we see three.

D

B

And they're already got a solid ideas. We are going here too much. Then you knows Philip here and evening time. Let's see what else we have in Cuba. We have a great comment that you're going to see you later along with reset, but we have coffee comment and you can sense. You see how a basic configuration file looks like, but so you don't have to remember it, but instead you can use the same command that is going to output basic equation, 5, but basic is probably not enough, so be extended.

B

That comment little bit and you have options to define API endpoint, to enable various features to coffee, good name and so on. So you can just run for the bike. Coupon offer training provider reached over, for example, and it will provide manifests. We can use footage thrush, but now so I don't have to explore all the class.

B

If you want just to export me full configuration for about features, there is full flag that is going to output this wall file and you can use it as a reference and we are going 12 put it like this. So, let's see what we have like in another file, we had a blabber Janet kind. Here we can, you find a casa name. It is a 28 in previous case use a time from which populated it and then we versions, and so what is the first thing you capable? It is class the network.

B

This is the coffee go some. It uses for constant services, service domain name, not port range, CNI plugin that you want to use. They said we have kennel and be met in case you can. Oh, you can't use the question, but in case effeminate this can be searched. True, you can recruit the connection cloud provider supportive, always our ABS digitalocean chatter office, tech, pocket vSphere, but in case your provider is not supported. You can just use none.

B

You will not be able to use called provided, specific integrations and terraform integrations, but you will be able to provision counter planes here. You can tell cubelets to configure for sol provider which occurs the point as a new CCM computers, manager or shorted CCM is responsible, so many Ron, you can get notes.

B

Ccr operates.

D

Ip Alice's.

B

And for some providers from each tool ocean, it is possible for allowing you to create services of dialogue balance in meanwhile, we can see that after running scale, we have got our machines. If we have free working modes is Val and besides compromise yourself, more interesting probably can be features, and this is to enable bar use features.

B

This is, for example, post acuity policy by default. We before Arabic rule to auto in school system namespace, and then you listen up to the priority that is needed for Cuban. It is, then you can configure it. However, you want by default. It is disabled, but you can just Tory note by saying in a bookshop.

B

We also can configure cluster to work with dynamically clocks. After this is enabled you require you to create a link object, and here is a link to documentation to learn more about works similar to prosecuted policy. It was enabled by enable true metric server is enabled by default, and it allows you to use comments such as you cuddle top and beside that. We also support configuring Cuban. It is to work with top and ie connected. Of course we don't if we don't deploy, operate ID provider.

B

This is responsibility of the operator, but this allows you to configure communities to use existing provider.

B

As we said, using it agrees such as their form is optional, and you can just use the host key to provide installation where you want to deploy control plane. You need to provide public on this optionally, private others. If you have private at the interface SH word as a username, and if you have a specific age associated private key file, you can define that he'll use it to connect to nodes for Easter, binaries front you baby, n, and so on.

B

Beside that we have API endpoint, as mention it in the most of cases. You want to use a log balancer here, but if you don't here, we can resort to using some of the nodes which is not recommended but can be use it to get manic.

B

These defines that we are going to enable motion controller. It is the point by default. If provider is supported by coupon and Mishimoto also, it is possible to use proxy for daca diamond and cubelet if you're on Akina environment that needs it, but just like you can define hosts, you can also define worker nodes in this file. So you need to provide name number of replicas and they provide the spec is define.

B

What is how is it going to be created, like Mei ability, zone Reggie on this size operating system and more now, to demonstrate it? It's actually what that you can get the classic without tire format, anything we are going to use such config, which says D, which is going to create coupon webinar faster, and this time we are going to use lower kubernetes version. So we can demo say that we can upgrade for 125 to 142 or NamUs own, and, as we said here, we have free cotterpin.

B

Is this defined as well as mobile to the user, this API endpoint and we define it Vaughn machine deployment that is going to be created with free replicas?

B

It has public key, so we can optionally asset and information about infrastructure and all needed pieces to connect it now, if, in this case, we're just going to create in free of abilities own, but you will probably want to have actually we can do. It is try not this. That's.

B

B

So, ok, let's proceed now whatever UI first to do. We can see that we've got our work, notes and you're going to reset the cost.

B

So what is that doing? It is a rotating work done by truebud and we are going to provide the partial views before there are four state as well as we are going to remove binaries as we are installing all the kubernetes version. This settlement is very useful, like if you want to experiment enable some other feature. You did not enable last time or you want to upgrade upgrade for up.

B

We usually use the super Napoli comment, but if you want like, in our case to go to water version, you need to reset the classroom and start again in case you. Do that remove binary secret by default? Will you buy no risk? But if you try to install rubber version, my stuff may happen because cumulant is the newest one and it can break some stuff there going to wait. A few moments. I.

B

Here we can watch our nose disappear. We can see that the first one is not ready.

B

Yeah now all three are not ready.

B

It takes depending on the API speed and so stuff minute or two so you're hoping that is going to be last one.

B

We are expecting for all three to disappear soon,.

B

Okay, already disappeared and what one is probably going to go soon.

B

Okay, it is done now. We are ready to obey diem reset to reset the Cuban I, discuss ER and if you for some time citrus rail yeah it failed, and now we are removing kubernetes binaries, and now that is ready. We can start again. Whatever you want to do is to run installed, but this time we need full manifest. You visited here.

B

Now this is going to take two or three minutes until we get a cluster.

B

At this stage, PDF of request files are generated, employed on all complain, instances and we created certificates. Cyclonus then, and now we are running qadian.

B

So this stage we finish it initializing the first corner, pin node and we only take about 30 seconds to ensure that all control, a cough components are up, and after that period we are training. Another caught by node then wait again at the joint third one and proceed to deploy CNI machine code over and more.

B

Now that thirty seconds passed now we are joining another cosmonaut car. There is no feedback that joining started, but this sub TPA is going to be fixed in the next version.

B

Okay, it is on it we again meeting 40 seconds.

B

Let's see, can you ready, yes, Austin? What, if you mother, why didn't you keep comfort I'll.

B

Let me eat this time into joining third and verse counterplay note, and it should be done in next few seconds.

B

And it is done, we probably got your profile, he has been committed, we can watch you come dead notes. Two holes are already ready. The first one is now ready. We deployed CNI an astonishing story, we are not waiting it to come up.

B

Okay, it is successful. We can be closer like previous time. It leaves some time for what conditions inside we can see. If you have free motion appointments, however, we shall let free replicas saying this.

D

One, but it is not only okay and we can see machines.

B

And we have five to them. One already got a sign in the IP address. Probably all four left I'm going to follow up soon. You can watch that.

D

B

But while this is running, let's take a look at the following, so we have this design 125 version of Cuban, a disc lusts, and we know that your test, one is 142 and we want to update how we can do it with you. One easy you just knows to use the cue ball, a great comment which just ok so takes by today, configuration file, adoption retired from state. If you start from digression and let's see how it works.

B

So what is happening when we run upgrade the cutter plain walls are graded in place. That means that over SSH we are first upgraded code needed binaries, like committee, a cube rate, your cuddle, then we are running cubed.

A

B

You great and after that we use, try and go like when we start to ensure that all compressor, up-to-date and if not, we update what is needed to be updated. Content is our upgrade a little bit differently because we use cost of API in machine code roller. We are rolling out the machine deployment which is going to add a new note and then to reiterate the previous one and then to add a new one, and there is previews until we are not running the latest version.

B

So let's try that so what do we have here.

C

B

We should wait some time wanted. The botanist appears.

C

Marco, when you was copying.

C

Configuration for water definition, you cook it every time the same subnet, so other worker nodes would not appear because they are subnets. Are it Z dependent? Also.

B

Well, when you are doing it all right, we only have the so if you have such situation, this is a very good time to seek avec le baguette. So you made a mistake in the manifest and you want to debug it. You have machines that are getting IP addresses and you want to see what is happening, how we can do it you can use to cuddle, describe and describe as frequency like this one.

B

Okay, so what's happening, it's promised. Automation it submit is the problem. We forgot to change that, but this that was it a does. How we can't bug that it. How we can easy find was a problem, so what you need to do is just to run described, machine and just find in events. What is the problem in this case? Decided?

B

It failed to create, because this habit is invalid, for this have abilities are now how we could repair that you can use to cuddle edit and then the Machine deployment object, which you later recreate the Machine object and create the appropriate ABS synthesis. Now that we are running out of the time slowly, we are going to skip that, but this was a good demo, how we can do it.

B

So if you see at the bottom or one is disappeared and we are going to upgrade.

B

So, as I said, we will provide the full coffee and we also say to upgrade machine deployments.

B

So it is carpet, it will just roll out the Machine point. Subject.

B

Now the upgrade process usually takes some time. So what's the problem here yet.

A

B

What we need to do is to change the cuban in this version to the numerable in manifest, and this is everything you need to do to say. I want to wait that work and, of course, the sun. You wanna wait.

B

D

B

Is doing what is needed to determine about is assist where we are running to find operating system course, name's, Jared, comedian files and so on, and we're running preflight checks or at this is what how we call it and what is they doing? It is very find that it is safely to update and away to death version. It is running our remote in the running state. Are they have correctly involves?

B

Ip not IP address as much is what we have the manifest the vs q policy is expected, which means that we can skip one or two versions, and it is active defined that you have to admit only to the next version. If you want away to two verses, we will have to go to next and then to next again. At this point, we upgraded kubernetes binary so that it was cuban DM q right and now we are waiting one minute after upgrading to a solid cube. It is a after date. One minute is over.

B

We are going to run cubed EMF gate make some more time. Then.

A

Start upgrading.

B

Settled and inferred cultivate businesses now this is usually going to take about five or ten minutes depends on the speed, then also the time needed to pull the containers, images and so on. We are going to use this time and I'm going to hand it over to Artyom to go ahead to some questions. If you have any questions, please leave them bottom in the Q&A box and we are going to eat now.

C

Seems like you described everything very good, because there are no questions.

B

B

D

B

Canto pain, hood is an ik 142 already. Oh.

C

I see one question, so what's the difference between machine deployment and machines? So basically that's a cluster API abstraction. The way you deploy it, the analogy, the closest analogy would be deployments there's two spots, so there are three controllers that actually create with those objects well work on those subjects. A deployment object will create machine deployment, object, will create machines, take set machine, set, object and, depending on upgrading for rolling upgrade of machines, it will recreate and.

D

C

Like gracefully, will tear down the old machine set and create an increased amount of replicas for new machine set in case of upgrade and then actual end object, which is machine, is handled by machine controller, which is talking directly to cluster sorry cloud, API provider and basically asking to create actual VMs and, of course make sure they are provisioned.

C

So that's that's the difference.

C

B

Can always was probably because load balancer said traffic to the quarter, Plano. That is upgrading right now.

B

And it is back, we can see that we have three masters now. Only two are running. One point four, two point two, but SUNY yeah now Q. What is a graded on third one and we see 142, but it will take some time until Q, medium grade get one's internal components, such a scheduler, cuckoo tour manager and more yes, wait.

C

And while I were waiting, I will continue with answers. So the question is what about ver metal and on-premise thoughts without provider.

C

Yeah, so basically, what you want keep one will only help you in this. With with in such environment, only with installing and initializing control, plane, nodes worker knows could be joined, using simple cube, ADM join command, which you can grab right from control, plane and.

C

D

C

Query create it will print your join command.

C

Another question is some machines are nothing but the cluster VMs yeah exactly it's.

C

Magic of course, it's VMs on any cloud provider. That's basically for Amazon instances, for example, oral digitalocean.

D

C

Is the same like hurt her talkin version for, unfortunately, I, don't know what what's happening in OpenShift version for so I cannot answer the question.

B

Okay, so at this time we have upgraded everything to resurrected, and then we presented to a blended conference, in our case machine controller, and then we I'm getting machine deployment option. So if you see and run your color get missing.

B

You're going to see office one, but if you describe it.

B

We are going to see that hewlett version is change it to 142. What you can do is to get pushing sets this art compasionate indication.

B

Hasari pipe works very similar to like deployments world. We have machine set similar to replica sets, and we can say that we have five of them so and here is form on the original one. We can see. That is thirty minutes old, NBC, poor bonbon, that is 62 seconds old, and this is one that is going to replace the first one on snow strain and it's a machine.

B

You can see that we have poor one, the old one and the new one with the IP address assigned. If you wait a little bit more, we are going to see problem with our joints and how it is going to replace. But now this can take several.

A

B

B

Okay, it is here we can rent it to get ready, and so this is going to replace first.

A

B

Scheduling the sailboat, so it is remained and so me it will be not ready and disappear.

B

Making wait a few more seconds to see this is going to happen now.

B

If you take a look at machine set, we can see that replica slot zero it got down skilled and after we disappeared, machine gets deleted. Machine set will be deleted as well and.

A

C

So let me answer next question: is there a support for big air gap? Installs?

C

Yes, a kind of at least you would need a proxy system which will allow to download all the images and binaries, but the proxy support. Is there.

C

B

Okay, so it is now not.

A

B

In a few seconds, we are expecting it to disappear from here as well and forward motion to get deleted.

B

If you have any more questions feel free to as well as mention it. There is a ternary box below and you can send your questions there. So.

C

Next question: while there is a difference and keeps Java versions,.

D

C

It's not cube city over from I suppose it's cubelet, we're friends, because we are currently I'm grading the system, that's why they are starting to change. Yes, we.

B

Started with the so we can demonstrate how upgrade process is working.

B

So this one is solving the keyword version, which was other so I've made every words.

D

C

Is it taking so long? That's because machine controller also drains the node using eviction artists, so your extra workhorse would would not suffer a lot at least and get rested out, another most gracefully respecting for disruption, budgets.

B

Now that we are running out of time, okay now it disappeared until is gone. We have only one working out and if you see machine, there is only one piece valid.

A

B

Machine set, it will be probably gone in next resets, as we are running out of the time. Let's go to our site and see what.

D

B

So most people today, a buddy so is q1 a tool for managing Cuba. It is class the lifecycle you can find it'll be hub on in cumulative organization, and we were coming any ideas. Services as well as contributors are very welcome. You can follow us on twitter, on Bootsy x, mo da kron4 EEG to follow us and also check out Rossi blog, where you can find a more detail it for coupon and we will be posting some articles about q1 in the future as well.

B

If you want to talk to us, you can join Cuba metrics like in Cuba channel. You can ask any questions, research, essence and so on. Thank you for your time. I. Thank you for joining us today. If you have some more questions for the end, I'm going to be here, three or four minutes for them. Thank you.

C

So this question about angling on google green deployments: that's probably about working with pods! So that's more of responsibility of cluster, a greater of cost o user. Even we only managed.

A

C

Nodes, not both I.

C

See some someone typing in chat.

C

D

B

D

It's it hub slide it with it. We have it's a github slide. We have slide yeah.

C

Well, this is one with all the links currently on the screen. Yes,.

C

Okay, people are looking at links.

C

D

C

C

That question thanks: everyone looks like we don't have new questions anymore.

C

So we can counted this webinar I.

A

Think looks like we got through all the questions then right at the top of the hour here. So thank you both for the great presentation and thank you to everybody who joined us today, just a reminder that the recording and slides will be online within 24 hours, and we look forward to seeing you at a future CNCs webinar thanks. Everyone.