Red Hat OpenShift OpenShift Coffee Break | Red Hat Livestreaming, 14 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OpenShift Coffee Break: Disconnected Series

Description

This is part of series of Production examples of running OpenShift on restricted networks, sharing experiences and best practices. We'll talk about OpenShift Disconnected installation on VMWare vSphere with IPI with Robert Bohne.

A

Good morning, everyone, my name, is jafar. Thank you for joining us today, so um we will be. uh I will be your host today for the openshift disconnected series.

A

We have our esteemed guests, robert and terrell, who will be speaking about.

A

Openshift disconnected installation on vmware this year, so we are very happy to have you uh on board today. Tyro will be the main speaker today for handling the robert's presentation. So uh guys, can you please uh introduce yourselves.

B

Okay, hi, I'm terra, based in in the north, where spring is coming maybe summer is coming robert. To be honest, robert is going to do with all the fun stuff. I will be just uh taking care of that. All the questions will be asked and uh so and so forth. It's a really cool topic. I would say: disconnect disconnected installations, are they are pretty enterprise credit, since disconnected environments are pretty common and we discussed that there is always some breaking news in the id sector we discuss what might be today.

B

I think that safari is from paris and we have colleagues that are from paris that I think that it's breaking news that psg went forward in the champions league, because that's a big thing for the sector, because madhu is much happier when psg go plays well and yeah he's more effective, then yeah.

A

Productive yeah way way more.

B

Productive he's not crying and everything so robert.

C

Yeah hello good morning, my name is based in the south. No, it's not really south. In the europe context. It's a middle! It's south of germany, um we're waiting for spring too, but it's snowing outside it's it's it's weird and yes, I have prepared a little environment on vmware and to play around with disconnected installation from installation to operator.

C

If you have questions feel free, we are in available via chat. I can see I cannot answer in the chat because I have not looked into twitch, that's funny. um First of all, I want to talk a little bit about um what is disconnected what is restricted on this aircap, because um both different wordings are flying around.

C

First of all, in our documentation we mentioned uh restricted network. So if you're looking into the openshift documentation and you're looking for some kind of disconnected air gapped, you have to search for the world restricted.

C

um Technically, there is a little bit of difference. So disconnected means you have a cluster that have no direct internet access, be blocked by firewall rules and may have or may not have a proxy so that you can slightly connect to some kind of internet through a disconnected thing.

C

Restricted means at the end. Are your network have or your cluster? That is some access to some kind of content, and so it's it's half connected how disconnected and air gap means completely accurate, so this think of military installation. So you put all your stuff what you need on a on a usb stick or burn it on to a dvd or a blu-ray.

C

It depends on the amount of data, and then you drive to your data center or to your aircap location, and then you copy all your bits from your pocket device into the data center. That is basically agaped.

C

We use quite often they were disconnected, but yeah in the documentation is restricted with the air gapped roading I described basically the procedure. Let me share my screen. A bit um hope it works share.

C

Hope you can see my screen perfect cool. So basically you have two different side. You have a connected side where you have internet access and oh yeah, and you have another side, that's your disconnected site or your aircab site, or your site restricted network just for for drawing, and basically you have to copy all your content.

C

You have to. You need a connected site where you can download all the content, and then you have to move your content to your disconnected location for a disconnected or restricted network. It's quite easily. You need to kind of jump post or machine and box. Where have access to all the content you need, and then you have to copy to your restricted network in the aircraft environment. Of course you need your pocket device and some time to travel. We all know. Traveling is quite hard today and technically on both sides.

C

You have a different um tooling on the connected side, where you can download your stuff or you can copy your images. You have various tools, you should all know very well the oc command line tool and with the openshift client command line. 2. You have some mechanism in to mirror all the content of an openshift release for the operators, or at least only some images, maybe for your app deployment or you use simply scope here.

C

um Scorpios is a tool.

B

One question to clarify: yeah, of course, some people might not understand what the content actually means in this case,.

C

The content- and actually this means uh case means um container images, so all right, yeah to install openshift. I you only need a bunch of container images and in the.

D

C

World another world, uh your operating system called uh red hat, interesting scores; okay,.

A

So it's a it's in ova or what what.

C

uh In the in context of vmware, it's an ova.

C

Let me show you a little picture, maybe you you have seen this one in a couple of redhead presentations, so that's basically the um process how it disconnected and stacked installation works. So you have to mirror all your content to a local registry. What is here really important and the registry have to support the docker version, 2.2 spec, but all registries on the market support it. So I have never seen any registry, there doesn't support uh 2.2 and then you have to copy all your content.

C

Important is also you need an um for the initial installation of your cluster. You need this redhead core as ova available on the web server, and this ova is during the installation uploaded to your recent. So your disconnected cluster. We talked about vsphere, ipi installation, so fully automated. You only run one command and then your openshift is running on um on vmware.

C

You need access, of course, the dv center. So the whole time. Also, if you have a disconnected environment or air gap, you need access to your vcenter if you have an integrated setup or and yeah fully connected to the to the vm environment.

C

Just this theory: um let's go back a little bit on this disconnected side on your cluster. You have basically uh two mechanisms to um point your cluster on your mirrored images.

C

So um imagine all your cluster is using creator.io, slash, openshift, slash some images, but this doesn't work anymore because creator.io is not a waiter when you're disconnected environment, you have your own registry and to rewrite those image addresses from creator.io to your own registry.

C

We have an object, that's called image content, source policies, and this describes how to rewrite your um image urls to your local registries. So you want to start a part with an image called um creator. Openshift blah blah blah and your image content policies says no, don't go to creator. Io, that's not available! You go on your local registry and pull the image from there, and this.

A

C

Image content source policy.

A

So robert uh yeah can we pause here just once again and to to explain uh what type of uh container images we need and why we need that for the installation. uh So I I believe that openshift uses what we call operators and they have to all pull those images before uh they get exactly uh deployed. Yeah.

C

I I will talk about this in a couple of seconds. I just want to clarify the basics.

D

C

Is important for the for the installation to understand what an image content source policy uh does and what it means? Okay, um just to have a full understanding. Of course you can adjust your deployments on a disconnected side to use direct your um your mirror registry, but this is quite hard for, of course, you mentioned it, the operators and the core open shift.

C

So let's talk first of all about the core openshift concept, so um we talked later about the operator hub and how the operator, but first of all, about the core openshift.

C

uh Maybe it makes sense to to switch to a terminal, uh whereas my terminal it's gone, then let us start one and I'll.

D

C

Can I remove this here, this thingy zoom yeah got it.

C

For my environment, I have an is it this cr, this screen size, it's not big enough right, isn't it.

A

uh Yeah now it's fine! It's fine, okay, cool.

C

A

C

This is um and hose that is connected to the connected world and is connected to and disconnected environment. Our disconnected environment is.

C

Basically, an vmware environment, of course we talked about vmware and it is a vsphere 7., and I have here a bunch of clusters running not all uh disconnected um a couple of ones are connected and I played as with some windows containers, but this is maybe a topic for another session. I have a few two classes: infra and demo.

C

There are disconnected clusters, and now we can start to install a second cluster, a third cluster at least the mode two.

A

We start robert yeah. Could you please uh zoom in just a little bit on the oh? Thank you.

A

It's sebastian, I know, he's getting a little old, so he doesn't see you very well, but it's a good reminder.

C

No very good, thank you. I have got this quite often um all right, so what we have to do to install an openshift um into and disconnected or at the end environment. First of all, we have to mirror a bunch of images. This is to be honest, quite easily. I have here some some notes.

C

I have done this before, because I have to prepare something um with the command oc, adm release. You mirror all the images we can start this again. Let's open this terminal, you can close twice the same.

C

Oh here, everything's fine,.

C

So side by side it makes easier.

C

So um and of course we have query registry running create.example.com is our registry in our disconnected environment- and here we can see, I have mirrored a bunch of images. An operator and important part is here this um infra openshift 4 registry or yeah registry. With this couple of openshift cluster or openshift core images, usually they are attacked with the version at the beginning and then with the architecture and the function. What the image is actually contains. I have two versions: mirrored version, 4.0 dot.

C

And 4.7.2 um they are available just just for um demo purpose. We rerun this. um Just again, we have a couple of environment variables. We have to set that's all documented in our documentation and then we have this command called oc, adm, release, image, oc atm release, mirror jesus um and this command copies all the content. What you need for core openshift installation from creator.io from the red hat registry into our own registry, and if we uh copy paste it and rerun it, um I have an arrow in my command uh jesus.

C

B

This it's not exporting the environmental variables, it's not recorded, run source and without aids. Oh yeah, you're right!

B

Thank you. Tara. I tested this this morning.

C

um The mirroring should be quite fast because it actually doesn't mirror right now. It checks on the redhead side, hey which image they have to mirror, and then it tries to push it on create.example.com, but the images are there. um During this run. We have here and pull secret um mentioned. This is document how you create one, because you need credentials to get the openshift images from the red head registry and of course, you might need some credentials on your local registry in myronqueryxample.com.

C

Everything is added to the pull secret. I don't want to show this pull secret right now, um but trust me there are credentials in it and at the end you get this. um I mentioned earlier this image content source policy, and this is important um you have to to save this. You have to store this somewhere in the text editor or whatever, and this is exactly the rewriting. So if your openshift core platform tries to pull something from creator, I o slash openshift, release miners blah blah.

C

Then it has no go on a mirror called a create.example.com infra openshift and same for the for the other images here on ocp minos. V4 are def, don't ask me why there are the names, the names um that's a question for our engineers, and this is actually the rewriting configuration we told to the cluster. So now we have to mirror our images. We can install um our openshift um for an openshift installation. We need an install config. um I will zoom a little bit in here on the right side. ah This is zoom seeding.

C

Oh, that's! That's quite huge. You may know this insta config, so you describe your environment. You can create those install config with open shift, install, create, install config and then you can run through this dialog vsphere audio credentials, and then you get basically an normal install configuration. Then you have to adjust those install configures created instaconfig, here's a in the browser window. On the left side, I removed a couple of credentials here: our vcenter sv center.example.com.

C

um We installed everything into a folder and blah blah blah here and yellow. There are those parts they are important. We add here this cluster os image url. This is an url to a web server where we have downloaded the red head cores image and added this sha at the end.

C

So, basically we are going to um to our mirror openshift.com.

C

um We are looking for the reddit course the latest one 4.70, and then we basically download this ova put it on web server on your disconnected site. We grab this um this jar from the text or you create your own with your command line tool, and you have to add this here. This is important because during the installation, the openshift insta command and binary download the ova and edit those ova to your vmware environment.

C

The second part is the image content source policy. So actually our image url rewriting configuration.

C

Then, of course we need still need a pull secret um because our internal registrygrade.example.com as an authentication and last but not least, we add, and certificate we add in root certificates on root ca as a to the additional trust bundle, because our created example.com registry has a certificate with what is not signed from a public authority. It's signed by an own authority and we have to add this root certificate to our additional trust model, and then we can run this openshift install command. uh Let me prepare a.

D

C

Copy the um prepared, install config into the directory.

C

Unfortunately, I cannot show you this configuration right now, but trust me it's basically the same, but I don't want to show this vcenter credentials uh into the internet, then create cluster minus minus team. O2 prey to the demo guard hit the enter button.

C

C

And then the installation starts- and here you can see it- tries to download the reddit course image from our own mirror. It found some cache hit in our home directory because I run several installations and then it creates. We see all the resources we can go into the vmware environment and uh but I hope we can see something here.

B

Hey robert one, quick question: uh yeah: how do you define which version of openshift you install.

C

This is defined by the.

C

Version of the openshift install binary. um If you run openshift uh install version, then you can see um which version is installed in our. um It is version. Seven.

B

Zero two see yeah that's kind of important. I was just thinking that if you have an openshift install binary for six and you think images for four seven, it will fail.

C

Yeah, of course, of course, I guess uh you're asking good question. um If I remember correctly uh to be honest, you have to extract this um openshift install command from your mirror registry.

C

Let us check the documentation, it's always good to to learn how you can use the documentation. You go to installation, you go on installing on vsphere, and here we cannot look for disconnected. We cannot look for airgap. We look for restricted network.

C

Network customization blah blah blah classic installing on a cluster installing a cluster on a in a restricted network, and here everything is um well described and all the what you need the prerequisites, some dns entries. We have everything prepared, um also adding vcenter root cs to your environment, because maybe your vcenter have a special root ca.

C

um This is all prepared by my own and create the inside conflict, no there's no extraction anymore. That was, I guess in the past. They also mentioned this cluster os image. How to add the pull secret, the additional trust bundle, uh the image content source, so the rewriting at the end.

C

You create a backup of your insta config, that's always a good idea um and again a bunch of documentation, and there should be also um this creator requesting a mirror host.

C

Where was this uh release mirror commands release mirror? Oh, I assume this was.

A

Yeah yeah so uh robert you're, looking for just uh to see if I can help uh you're looking for the comment to extract like all the images that are uh needed for a specific release,.

C

No that one here I was looking for um this year is also important. To be honest, I forgot this. It works too, but at the end you should run this.

A

Because you have the oc release environment variable where you set the version like the ocp environment, variable.

C

D

C

Know manifest that's it's very funny.

C

Infra openshift 4.7.

C

But this is a local secret.

C

God of demos and let's check our registry um open shift, and then we need to tag for ah okay.

C

The architecture is.

D

C

D4, then, you have the extracted version of your openshift install command, and then you have also this release image pointing to your internal registry. um To be honest, I forgot this step.

C

It works with the normal openshift insert command as a variable, but this step ensures that your openshift install command version exactly match the version to your images. You have synced, and this is the fully answered to you right at the uh tell about your version, questions yeah.

B

C

You're welcome. um I forgot this well um funny. um Our installation is running through um it takes a while. uh We have access to the cluster um at the end for um operators.

C

Let's go maybe should back to to our presentation for the operators. You have basically have done the same. You have everything copied to your disconnected environment, um but for operators, you have um two things to discover and you have, on the one hand, side all the operator metadata or at the end, um what your cluster, let's log in to one of my disconnected clusters, um we have this operator hub. Oh yeah, I lost my session.

C

Operators operator and you have to sync all the information, what your operators are available in your disconnected environment. So, basically, technically you have to sync those tiles here.

C

This is the first step and the second step is, you have to sync all necessary images that the operators need to run.

C

And then it should work.

C

So basically, let's go back to my um kind of speaker notes um about the operator. Sync.

C

C

Yeah open shift installation operators um this little bit more. um First of all, you have to run and from red hat, prepared operator index image. um We only x have an example here about the uh rated operators, so you have to run the image and then you have access to the whole catalog.

C

Let's do this or maybe it makes sense to yeah. Let's do short so now we have running basically a back-end for the operator, and now we can ask which package are into the catalog, with a grpc curl.

C

A

We get the list.

C

Of all operators that are available in this catalog and then we can pick some operators we want, and this we can do with the opm index pro the opm tool is available also on mirror.reta.com.

C

If you pick your um choose the right version for the 7.0, for example, and then you have here the download link to the opm tool- and I assume it's again too small um here- the opm download link.

C

um With this opm index prune, you basically pull the the retina catalog image. You say with the minus p a parameter. You only want the this op, um this advanced cluster manager. um Oh it's duplicated! If you did, let's show the help um which package you want.

C

um You run it let's, let's do it so we select from the red catalog. Only the advanced cluster manager opens the pipelines operator and the web terminal, and then it's um the opm tool downloading the image clean everything up, and then you have an image tagged to your local registry with your operators you want to run.

C

Then you have to push this catalog image to your registry and then it is a way to integrate into your registry and then now the important part is you have to mirror all your images from the operators that you have picked with the opm tool into your local registry, and for that you have to command oc. Adm catalog mirror- and here is important- to um run or to sync every operating system so keep in mind.

C

Openshift is available on various architectures, um basically 864 on ibm set and then ibm power, and you have to sync all architectures um here, for example, I added also the manifest only thing, so I basically do not sync at the moment everything I just create some manifests and again some.

C

ah I forgot to export this credentials.

C

If I run this now, then I have this kind of directory here oops. This was.

C

Sorry from directory, then I get this manifest directory, and here I have a bunch of files. um Let's make this a bit big. We have, on the one hand, side this catalog source. You have applied this to your cluster.

C

Did this says to your operator hub? Hey, you have a new operator source, please um go and look into that image. This is your new backend for all your tiles in your operator, and then we have as an image content source policy.

C

This you might know what it is. This is a bunch of image rewritings and you have applied to your cluster too, um to rewrite here, for example, for the pipeline operator. So if the pipeline operator tries to start an image from retrospect.io openshift pipeline stage, blah blah blah, then please go on slash: create.example.com pipelines, blah blah blah. So because registrywrite.org is not available, you have to apply this, then you can run oc.

C

Image mirror command with this mappings txt as important is to um add. This parameter skip multiple scopes. um True, um because with quake we have some rate limits and some um yeah enterprise features, and then you have to tell the oc command for the image mirroring. Hey, please careful with this registry.

C

If you send two big http requests or too many requests, then create a blog. You- um and you can avoid this. If you add this, this command line switch here and of course, you have to filter all the the architectures and the mapping. Txt is also here and the mapping txt files basically and um every line is an image, and here we take look into the last here. It's bigger and the last line, and it says: hey copy from red head registrator.

C

Oh blah blah blah this to um to create.example.com, and here you see an important thing. We worked in the operator world and with openshift itself. We work with this digests and this is quite important to know our um our content source policy mapping only works with digests, so we have here this content, image, content, source policies, and this doesn't work with tags. So you have to add your images with digests.

C

If you start in pot with red registry, I o recom tunnels, rel and then you add in tag it doesn't rewrite to that mirror. You have to add and digest at the end- and this is quite important to know, um and this is why we have acer to sync all um architectures, because all operators does uh have the um this multi-arch image digest um in the deployment config and if you're, only using one operating system or one architecture support you change the digest at the end.

C

That's why you have to sync the whole image um with all uh the whole manifest list with all architectures at the end, and then uh you take care that the digest of the money of the multi-argument stays the same um hope. That is clear for you, that's quite complicated.

C

um I have a blog post who describes very well how this works. um Yeah.

A

If you have the the the link, uh please yeah, is it on openshift.com.

C

uh No, maybe we can republish it on openshift.com. um It.

D

C

In my um notes, jafar, I guess shared the link into the chat yes and on the top there's some documentation, resources and here's an a link at the end continuum which is multi-arch manifest ids digest. What's behind, okay,.

A

Let me share that on the chat.

B

Yeah, thank you one question. uh Can you add because you don't probably in the beginning, you don't know what operators the developers want to use. Can you add those later on to the cluster, so you don't have to in the beginning, know everything.

C

So, basically, those um steps here for the operator printing syncing uh can have done through the whole life cycle on your cluster. So it doesn't matter, it's not have to be done quite after the installation, as so after a year. You you recognize. Oh, we need this uh kafka operator. For example, then you can run those steps again. um Sync, as the kafka operator apply everything to the cluster and then your kafka operator is available.

C

This is also the same procedure. If you want a new version of an operator, you have to resync all the content from the new version of the operator to your environment, then your the new operator version is available and, of course, for your application developer. They have also to sync their own application. Images in the disconnected environment.

C

Yes, we have applied everything to the cluster and let me yeah- and now we have some a couple of um post install configuration. We should have done to your cluster. um For example, we have to configure our registry to add our own registry to the search path, a bunch of developers or.

C

Don't write the whole name of an image, so the full qualified name of an image into the pod definition so that don't write creator. I o slash blah blah blah slash my app. They only fright my app as an image name and then, if the part is scheduled to a note, the note have in search paths which registries are into which retrospect we are looking for the image name um and this we have to adjust a bit.

C

And for this we have this image configuration and we have to add our own registry, our mirror registry, to the allowed registries for import. This is important for the image streams and as a registry source, and here is important- the container runtime search registries.

C

We have to add our own um mirror registry and of course, we have to add the allow registries so that we allow images from example.redhead.com from create.example.com our own mirror registry. And, of course this is. This is important to know. We have also add uh registryread.io create.io registry access retro.com. Why?

C

If you do not allow images from creator.io or the registry.io, the openshift cluster tries to start an import, a reference to the registry and then the um cryo environment. On the note is: oh, no, it's not allowed to start an image from retrofit.redhead.io.

C

It's only allowed to fetch images from creator.example.com, because this rewriting is, after the filtering, so first the environment filters which graduates we are allowed to pull and then um the rewriting takes in place. So um this is this is the thing you have to know and it's documented into our documentation, and this is also the same for image streams.

C

So if you are using um image streams in your cluster and you run, do you want to run um oc tag or oc import? Then you have to allow your mirror registry and you have to allow create an arrow. If not your cluster upgrade will fail, because during a cluster upgrade it recreates some image stream and image stream text, and then it runs into an arrow. I documented this arrow here. um It cannot um allow.

C

It cannot add the the image stream tag because it's not allowed created or to import to your own um via image stream. So there are some pitfalls. You have to take care not, and um you want to disable your default sources in your operator hub. So by default, um your operator hub have some default sources.

C

um It tries to connect to the internet because the internet doesn't isn't available in a disconnected environment. Then it's important to disable all the default sources for your operator, because you add your own sources with your own operator images.

C

Some other things, um configuring, your image registry and maybe add some ntp settings as documented here is just just for me. This is my personal notes. I just share this with you. um So let's do a little bit recap. We have the openshift installation. We have to mirror the release images. We have to mirror the red head cores. We have to prepare the install config to add the image content source at the ca.

C

Add this cluster os image, and now I want to show you something really helpful. If you don't know what your config options are readable in the install config and you are on your disconnected side and you don't have access to the internet, you can run oc, uh you can run openshift in style.

A

Yeah there's just one thing that we maybe forgot to mention ahead is where we get like the binaries, uh the openshift install, uh and I will send the link to the openshift. Try url, where we can see the different options, uh then they pull the installer from there and then they start. This workflow.

C

Exactly exactly right and if you're on a disconnected side, you have to copy your opendrive inside command, your oc client, all your images and you forgot the exactly setting in your insta config. You can run openshift, install, explain, install config dot. For example, platform.

C

You see all the platform options we want to know about vsphere. If I can type vsphere, then you see all the vsphere options, and here you can see.

C

Where is it cluster.

D

C

um Cluster os image overviewsql provide on redis web about to download the retina score ova. So um a documentation is always um with you if you have a command line and you have access to the tools just for information. That is the same for the oc tool, but you have to access to a cluster.

C

The installation is running. We have done the post installation steps, so we have configured our registry. We have the operator hub. We have for the operator disable the default catalog source. We have created our index that we have our tires in our operator hub. We have to mirror our operator images, we have to apply the operator source and the image content source policy for all the rewriting, and now we have a cluster with um with some operators available. Maybe we have we have some time. Is there some questions left.

D

B

A

Yeah, I don't see any questions on the chat for the moment. Good.

C

um I have here um demo one cluster that is fairly empty and if you like, we can apply all those um for the operator hub. We can run through this mechanism to apply this catalog source, blah blah blah, and then we can run in pipeline. For example, we can also talk about and how to run an upgrade. What do you think what is necessary.

B

I think increasing the font might be registered so increasing the font.

C

Oh from from terminal.

B

Or I think both yeah.

C

This is not a big deal. Thank you.

D

C

Question it's great to see deployment.

A

Okay, so I uh seriously have a department.

A

um There's a question, but I'm not sure I understand it.

C

Yeah, I I don't know which, which deployment is mean deployment of a cluster or deployment of an application.

A

Yeah, let's just give.

B

A giveaway, I think that it's a deployment of the cluster, because it's referencing upi and ipi in there and and then I think that once uh we did all the steps, the installation is exactly the same as with uh internet connection.

A

Yeah, so maybe just to clarify uh what we are dealing with here. So, um as you might all know, openshift offers different types of installation. We have what we call the ipi, which is a fully automated installation that uses the installer, and we are also uh starting to provide the graphical standards that we call the assisted installer, which is in in preview now and basically, when you go with ipi.

A

So what we call infrastructure, provisioned installer, it will take care of talking to the back end uh like the target deployment uh environment and it will create all resources like in vmware. It's going to create the virtual machines. It's going to uh bootstrap everything in aws. It's going to create uh the vms. It's going to create the networking.

A

D

A

Ingress the load, balancers, etc. So what we are showing here is how we can do the same, but in a disconnected environment and because, as robert explains when we do this type of installation, we need to pull content from the red hat registries and because we can't do that directly because we have no network access. All the prereq tasks. That robert showed is just how you can, instead of talking to red hat, how you can talk to a local content store.

A

So for the images uh for the operator, images for the operator hub content and, basically that's it once you have that prereqs that is set up, the rest of the installation is the same as you would do with the traditional ipi installed.

A

I believe uh yeah. Let me check uh if I can so I'm talking now with one of the opportunity product managers, while you guys do something else and check if I can mention something that we are working on.

C

C

Trying to login into a demo 2 cluster demo, one sorry um cube admin.

B

You took the demo to password. Oh.

C

B

So good, I'm useful, of course,.

C

It's always easier if you are not typing.

C

B

C

Yeah, um exactly if you don't have yeah, if you don't trust the environment on customer side. I had this as a couple of times where lucky internet connections or strange things happen with proxies between your registry.

C

It could be easier to uh to run a disconnect installation, but um you have to also keep in mind that no, not all operators support to run in this connected environment. So um this is what you have to know. Basically, um all upstream operators are not prepared to run in this connected environment.

C

To be honest, so I have never seen one or one upstream operator what is running a disconnect environment and also for customers, it's quite hard to not only for customers for everyone who starts to learn and try something on an openshift or kubernetes cluster who's. Disconnected it's quite hard because uh you have problems, maybe with examples from an internet. You want to run an example from the internet. The image is not available. You have to mirror the images from your docker hub or whatever to your local registry, and then you have to to run.

C

um Those images so connected is to is useful if it's possible.

A

So guys well yeah. While we were talking about that, I you were looking for some breaking news.

A

uh I might have one uh today, uh so um I was sticking with one of the open product managers and he said that it's safe to say that we are working on automating, all of those prorecs, even more and providing like a single tool that would encapsulate all of those prereqs in order to make it easier for you to just run like a few command lines uh with that single tool, and it will take care of doing all that prereq work for you, so nothing official of course, but we are also looking at providing an integrated registry to handle the offline or disconnected environment that will have all those automated things synced in.

A

So it's going to be even easier to to do all of that. So it's basically, we are taking the brain of robert and putting it into a binary.

C

It calls operator.

A

Exactly it's gonna, be the robert operator disconnected operator.

C

So in parallel, I tried to look in to twitch to have access to this chat right now, um yeah. Well, what do you think about? uh We can prepare this uh demo cluster demo, one cluster to, um for example, install the pipeline operator, or we can take a look on the infra cluster. There's a pipeline operator running.

C

um What do you think guys? What is what is the best valuable thing.

B

I think that might be better to how to prepare the empty cluster with the pipeline operator, because that is the more closely disconnected so how you enable an operator in a disconnected cluster.

C

Yeah: okay, fine, then that's good, um I'm here locked in into the demo one cluster, three masters, three notes: um let's go.

A

Can we first have a look at the operator hub to show what's in there.

C

Of course um the operator is empty, so um I will increase everything a bit um hope you can see this. um It's a cluster name demo, one the operator hub is empty and if we go to cluster settings, cluster operators, global configuration- sorry my fault global configuration operator hub- and here we can see um they are features, so they cannot connect to the certified operators, catalog, the community, rated marketplace and retail operators.

C

So, first of all we want to disable.

C

All these sources, and for that we have this oc, patched command operator cluster and then it disabled all the default sources.

C

This is straight from the documentation, so we can open the documentation and this a straight copy paste from our documentation.

A

So when we do that, what operator is like rewriting its configure.

C

Everything is disabled here um it. Basically, let's go to the yaml here.

C

Basically it added this configuration okay, perfect so and then all sources are disabled uh operator is still empty.

C

um Here I have a bunch of my catalog source. The image is mirrored all images are mirrored to grade.I o created example.com.

C

I don't want to waste time to mirror everything again. So trust me everything is mirrored, and then we apply um first of all, the catalog source.

C

What happens now.

C

The dom we go to cluster settings global configuration if you, if you hear a baby crying in the background, my little son is a little bit sick and is angry of being sick. So, uh and sometimes it's crying quite loud. So sorry for that.

A

No, no, we don't hear anything so cool, but I hope he gets well.

C

Yeah, it's getting better, um just angry to being sick, so I have call from there um we added this um catalog source with this oc apply command prepared by this um oc. Adm catalog mirror command, and we have here this uh retired operator. Maybe now we can see, we have some tyres how it works. In the background um we go back to cluster settings to show you again, we have here um say: hey, please, the end point is an image, so we created this image. We push it to our own registry. This is the endpoint.

C

What happens in the background?

C

We go to the um my openshift marketplace: namespace marketplace, and here we can see um our image.

C

Is started so this is our index image. If we go to the logs, we see basically the same what we have seen with this pop and run command. This is our image. What we have built and this database only contains our three operator we have selected the operator, connects to the to the port, the uh gprc and ask hey: what's the operator touch yeah, thank you.

C

What operators are readable and then it fetches all the information and shows the tires here.

C

If we ins try to install this pattern operator straightforward. To be honest, it will fail right now because it cannot pull the images. Let's go to the open shift.

C

Operators we go to parts, nothing.

A

Here yeah check the events.

C

D

Installing and.

C

Oh, no, it's it's! It's not.

D

D

C

We can check the events overview. My events.

D

C

Nope, which name space was it.

A

um For the installation uh yeah go back to the to the info.

C

Registry path, you see here within digest. This is the important piece it's.

A

C

Yeah, the basically the images are not available. Oh, it says it's installed.

C

I should see that if you're creating yeah.

A

If you go to view it here where you open.

C

Shift operators yeah.

A

So during the time keeper we have one minute left: um oh wow, yeah yeah time flies. Thank you. So, just just as a as a reminder. We have a lot of upcoming sessions on the red hat twitch channel, which is twitch.tv red hat openshift official I'm gonna send the link, and today the first show will be.

A

I think, scott mccarthy talking about ubi images, so make sure you guys can.

A

Check out those sessions, if you are interested and thank you very much- and we also have the our upcoming uh coffee break, which will be on may 5th, uh which will be basically an introduction to the kubecon uh mia event. So uh taro and I will be present again robert.

A

You would be more than welcome to join of course, and so I guess we can say that we'll see you uh soon uh on on may 5th and please, if you have any final question, let's take that otherwise we will be able to close the meeting any closing comments, uh taro or robert. So before that. Thank you very much. uh It was very uh great to have all those in-depth details, because this connected install is uh still one of the like, not so automated things, but we are working on that.

A

So thanks a lot for for for doing this session. Today, um cairo and robert.

C

Yeah, thank you. Sorry, um I have prepared much more one hour is not that much time. I have learned this in this session.

A

Yeah, but this is a series- this is a series, so we will invite you again. uh Let's just make sure we have defined that in the agenda and you will be more than welcome to come back.

C

Of course, thank you very much.

D

C

You have any questions, feel free also to reach out on twitter to me at robert bona um feel free to ask contact your local redhead sales guy.

B

Thanks everyone for joining bye, all.

A

Right, thank you very much have a good day. Bye.

D