Red Hat OpenShift OpenShift Commons | Red Hat Livestreaming, 8 Jun 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OCB: Simplifying Stateful Workloads in IBM Cloud and IBM Cloud Satellite

Description

Stateful container workloads are increasingly being used to run mission critical workloads. To ensure that stateful workloads can leverage flexible and highly available storage that meets their app requirements, a tightly integrated container-native storage solution is required.
With the integration of an OpenShift Container Storage add-on, you can now deploy an operator-based storage configuration to the Red Hat OpenShift on IBM Cloud clusters that run on infrastructure in your on-prem data center or in other cloud providers by using IBM Cloud Satellite. Support of IBM OCS add-on on IBM Cloud Satellite also includes the ability to create a OCS single configuration that can be deployed to multiple IBM Cloud Satellite OpenShift Clusters. The OCS supports both public and private clusters across IBM Cloud Classic and VPC infrastructure, and is integrated in to the hourly and monthly Pay-as-you-Go billing plans in IBM Cloud.

A

A

B

Hello, everyone welcome back to another openshift commons. My name is karina angel and I am an openshift product manager here at red hat and excited to have with us today sandy ahman. I hope I'm saying your name right, akash and sharisha from ibm and they're here to talk about openshift, container storage on ibm cloud and ibm cloud. Satellite and sandy. uh Could you introduce yourself some more as well as a caution sharisha for everybody.

C

Sure I'd be happy to uh I work in our container team in ibm cloud and satellite and we focus on uh the storage integration within uh our container service, both for kubernetes and openshift across those environments. So I'm the lead architect uh driving various storage integrations, including openshift, container storage. An idea, uh akash and srisha are part of our development team and uh I'm happy to have them also participate and demo. Today, the integration that we've done.

B

Awesome, thank you very excited to have you all here. I know you have a presentation prepared as well as some demos and let's get right to it. I'd love to hear more about ocs on ibm.

C

Cloud I'd be happy to okay, okay, so let's uh kind of talk about uh what we're going to discuss today, uh we're going to discuss our integration with openshift container storage within our public container service and our ibm cloud satellite environments and how that can help simplify uh stateful workload, deployment across those environments. So well, we'll give a little bit of context today on the particular subjects that we're going to be talking about today.

C

I'll give a context on the offerings that we have today just to give a context of what they are and what they provide. The user.

C

Then we'll drill down a little bit on how we're integrating openshift container storage into those offerings, including uh today's supported deployment topologies, and what we're thinking in the future to support as well and then we'll finally close with the demo uh in different areas, so be happy to kind of talk about those today and kind of get going. So, let's get started, uh let's get some context around our offerings today and ibm cloud and satellite.

C

As you all know, many companies and enterprises are already migrating and creating strategies to move to a cloud first infrastructure environment. In some cases, those customers are moving and developing new applications in a very cloud native application. Modernization approach, however, in some cases, there's some applications that cannot go to that kind of model and so they're not as suited to be on public cloud environments, and so they want to still have the same experience of a cloud and experience to deploy their app and manage their app.

C

But, however, they need to keep those applications on premise, whether it's for data solventary requirements or or whether it's just for uh you know the fact that it's like they want to control their costs and it's hard to move because they already have that running in their infrastructure. But they want to modernize how they actually are managing their applications from a cloud perspective.

C

So within ibm we have two offerings that provide not only a manage openshift offering, but also let you have offerings that give you the full services of ibm cloud in any dedicated environment, along with that environment. We're also providing the ability to install and configure and deploy open, shared container storage across uh both the public cloud and on any customer provided infrastructure. We'll talk more about those in detail in the next following charts.

C

So let's talk about our first, offering red hat openshift on ibm cloud. What that offering is is a fully managed, openshift ocp, offering it's available across at least 35 data centers and growing. Today it integrates with all the ibm services that are supported, such as ibm keep protect for keeping your key management uh key, safe im and also you know our object, storage uh systems and our native storage systems today, uh as well.

C

It's fully managed and fully uh configured across the globe with our sres and available with all the security certifications that enterprises need today, just a little bit of a drill down on what this well ibm does for you. We actually help you with the full life cycle management of deploying the cluster for ocp and openshift, and we actually deploy this in highly available manner. So we manage all the masters for you.

C

You don't have to worry about that, manage all the updates, patch updates, etc, and we give you control of the client data plane as part of the client data plan. You can customize certain aspects of it, including deploying things like openshift container storage, which we'll talk about, but other things you can definitely customize as well.

C

With ibm satellite that's another offering we have an ibm satellite, it brings the power of all the ibm cloud services, whether it's cloud packs or watson, or our databases, and including eventually, our third party services, that you can then subscribe in a subscription-based manner on any infrastructure. Why do you want to do that? One reason is, you might want to control your costs. Although you want the cloud experience, you probably might want to also leverage your current infrastructure investment.

C

So we allow you to take the same experience that you can see in ibm cloud and deploy that on any infrastructure you would require and have, and with that one of those services that we do fully manage is a red hat openshift to manage service just like in public cloud, and then with that.

C

We will also support uh different customizations within that in the same business and will provide the ability to also tooling to move applications back and forth between those two environments and also connect from uh from on-premise environment to ibm public services as well.

C

One of the popular applications that we'll try and show you today is also the fact that many customers of ours are using cloud pack for data as an example application and that application can run in the public cloud or in any other environment and we'll kind of demo that today slightly today as well, uh there's a little bit of a drill down on the satellite uh kind of a big block diagram.

C

uh What you'll see is a customer will create their own location at their own site. uh The requirement for the customer is to provide their own infrastructure after they provide their own infrastructure.

C

Our satellite control plane will take over and help you configure the location and actually create the clusters uh in those locations that are managing the highly available manager. Much like we do in the public cloud environment as well. The customer is then free to then deploy any services like icd ibm calls watson uh in within that environment, uh including their own uh custom apps as well, so they can actually choose where they deploy and choose how they deploy as well and what services they want to integrate within their environment.

C

Now, let's drill down a little bit on uh the actual container storage integration that we've done, not that we've. Given you some context under the cloud satellite and ibm cloud, openshift preference to put everything in context within ibm cloud. We provide a lot of native storage capabilities, as most cloud providers do, such as file block and object.

C

So we also are heavily invested in providing our customers a very flexible container native storage set of choices uh and the reason for those choices are the fact that, within a hybrid environment, many customers are trying to have the mobility to move their applications in any direction in any cloud environment.

C

So the ability to bring your own storage and take that with you is very critical required for customers and with these solutions they provide a very consistent approach in managing the life cycle, management of the volumes being able to do the actual backup, restore and also the stateful uh workload migration between those between clusters as well. So these types of solutions, such as red hat container storage, provide much more advanced capability than the basic native storage uh capabilities due today as well.

C

So it's a good choice if you're trying to work in many hybrid environments as well, and that's why we're offering that very tightly within ibm cloud along with other partners that we support in this space as well.

C

Let's take a a quick look at how we plan to offer ocs within our cloud environment, for both satellite and public cloud, we're going to be offering two packages uh and how we offer ocs essentials in advance. The difference between essentials and events is this: the essentials will provide the basic life cycle management of the volumes, including volume snapshots, all the things you can do through a csi driver.

C

The advanced package will provide things like more advanced security capabilities, ppop encryption with your programming, pvc encryption, but also multi-cluster, dr and backup capabilities. So we plan to offer these uh on a subscription basis on a and and based on, uh if you're, using a vsi or a vm versus bare metal, the pricing will be different. We're still working out the pricing, but uh that'll be forthcoming soon, uh once we do the ga.

C

Currently, we are in a tech preview mode, so it's available for free for anybody to try it right now in ibm cloud and satellite uh you're welcome to try that out. uh We support a deployment, topology called internal mode and that mode basically co-locates ocs with the ocb cluster and your workloads. As shown in this picture, we are considering later this year supporting the ability to deploy an ocp cluster and remotely connect a remote set cluster on premise, primarily for a lot of the on-premise side.

C

A lot of our customers may have existing ceph deployments on premise and we want them to be able to kind of leverage that and be able to connect that they still have the same experience with ocs and using that remotely as well. So that's uh one mode that we're also working to deploy and, in the very uh more longer-term future, we're looking to support a kind of a centralized uh storage, ocp cluster that runs on openshift itself and can then be connected uh by multiple ocp manage clusters, as shown in this slide ground.

C

The schedule for this is uh being planned for, but you know this is just the big picture of what I wanted to kind of show you the modes and topologies that uh are capable and we plan to support uh in the future as well.

C

So, let's take a quick uh drill down on ocs uh ocs brings together three projects within open source together self, uh nuba and rexef provides the file and block capabilities. Nuba provides the object. Capabilities brook was designed to provide an operator to help configure and manage some of the saf aspects as well.

C

As you can see, there's a lot of artifacts that get deployed uh various human sets etc, and with this integration, we've actually integrated.

C

The fact that we can now integrate with ibm classic infrastructure and vpc infrastructure or any satellite provided is for infrastructure with ocs. So any any place you can run satellite is any place. You can run openshift container storage, including uh on ibm public cloud itself. We've done the integration so that when you define public cloud or even satellite, you have the ability to use ibm cloud object, storage as you're kind of backing storage for anuba, uh and also for any uh future capabilities around uh multi-car, vr, backup and integration with backup, tooling, as well.

C

um One of the things that uh we have done taken time to do to integrate in ibm cloud is: we've tried to make the experience of integrating openshift container storage a little bit more easier and embedded into the actual managed openshift experience within ibm file.

C

uh Because today, if you were trying to install openshift container storage- and let's say you wanted to install that with local disk using localdis, you first, if you're going to do this from the command line, you first have to install the local shift. Local storage operator create all the name spaces, uh ensure you have all the disks uh all ready to go uh to feed the openshift container store, so they can get started to create all its resources.

C

So what we've done is created a ibm, ocs add-on within our environment, and we've actually tried to reduce the steps from you know as much as from eight steps to one step to install and configure it we'll actually demo this capability today with you today, but I'll cover it briefly and then the following charts as well.

C

So let's talk about the experience that you can get to see within our user interface you'll see once you create a ocp cluster uh you'll have an additional set of add-ons to deploy. You can add different capabilities. One of those capabilities is to install openshift container storage just to install it. You just click the install button and it's ready to go.

C

We even have some pre-defined defaults, so you don't even have to put as many defaults in to prime ocs they're they're ready to go, and once you do that you can actually target whether you want to deploy this on ibm cloud on vpc infrastructure or ibm classic infrastructure, either using remote disk or local disk and likewise on ibm cloud satellite environments. You can also deploy to uh cloud-provided infrastructures like aws, azure or on-prem systems, or your own kind of hardware that you already have around in your data center.

C

We've kind of really heavily focused on the ease of use in the life cycle. Management of this a little bit under the hood of this integration is our openshift add-on, consists of two components: add-on user interface and cli, and also an add-on operator. We'll talk more about this in the next couple of charts.

C

What this uh terminal window is showing you from the command line. I created the cluster in one shot, using the ibm command line, and what it's doing is it's creating an open ship, container's cluster and it's actually priming it with some parameters to actually feed it. The mod size and also the osd size, that's desired, uh and then, once you run that command it'll actually create the cluster using those parameters.

C

uh You can override these parameters. Otherwise, if you don't overwrite them, it will actually provide some defaults for you, so you're ready to go, and you also have the ability to auto create the cluster.

C

uh When you create this command line- or you can do that afterwards as well, we will talk about that as well, but you can see that with this one, when people are creating stateful workload, applications they're in a more of a dead test and production cycle, they'll have to kind of install and tear down these clusters in a repeatable fashion uh and having this integrated with the ibm command line is great for them. uh If they're on openshift on ibm cloud, they can simplify their cicd pipelines.

C

uh We also will have the ability to integrate with other services like ibm, hyper, protect and keep protect and in the future, and also specify the advancing essentials offerings through this kind of option, but also in the future. We'll also have the ability to just have t-shirt sizes and using these t-shirt sizes, you can then just give our users a simplified approach to say how big do you want your cluster to be and we'll go, create the cluster uh ahead of time with that t-shirt size?

C

So they don't have to understand all the nuances of these options like what is an osd size and what's the posty storage class name we'll do that all for you. So that's kind of uh what we're trying to do with this add-on as well.

C

uh Akash will cover this in this demo, but I just wanted to show you that uh our the ocs add-on operator that's used behind the scene of the cli that we showed it just takes a simple security and you can just you just need to prime it with a minimum of three inputs. You can give it more inputs, but there's a single point of control of basically uh having you configure all the inputs in one place without having to create namespaces and deploying different types of resources to get ocs going.

C

Yeah we'll cover more on this in detail, but it's a just one, simple crd and then you're ready to go.

C

You can take a similar crd that I showed you on the previous chart. This happens to be something that we would do for vpc environments, but you can use the same crd on uh aws uh and you can feed it. A a a class name and it would go, create the cluster uh just through our building, and we we assume some defaults. We actually always auto, enforce three-way replicas. So most of our users need that and want that for their stateful application workloads, all those choices are kind of taken care of as well.

C

You can have the ability to override what packing storage you want, whether it's cause by default or even use an s3 as well. When you go.

C

um The other thing that we've done in satellite environments is we've taken that operator that we've written and we've. uh We know that within a location, a customer will create multiple clusters more than one and when in those clusters, he's going to uh need the ability to deploy uh something like ocs multiple times. As a result, we've provided a grouping mechanism, so you can define a group and the group membership can be dynamic.

C

So, as members join the group, what will happen is with our satellite templating technology. You can deploy uh a storage configuration. We support. Multiple templates within satellite ocs is one of them, but we often we can also deploy uh aws csi driver, as shown in this chart, or even netapp or spectrum scale.

C

uh But once you do this uh anytime, a member joins a group. Ocs is auto deployed with a specific configuration, that's very site-specific, the site-specific credentials, etc and customization. So you don't have to do this over and over again, we'll demo that today as well, so um I provided a brief overview fairly quickly of what our our offering is on our integration.

C

We're going to spend some significant time on the demo next and then we'll open it up for questions, I'm going to start with the first demo, but before we do that, I want to show you what we're going to be demoing between akash and srisha, he's going to actually show you how to deploy ibm cloud ocs on vpc environments. It'll. Take you through that, and here she will show you how to deploy ocs on aws to two clusters in one shot, so that would be uh that's gonna be great.

C

uh Before I end uh my presentation, uh I want to do a quick, slight demo just to give the context so within ibm cloud you can see. We have different services, one of those services, the open shift services. I can actually create clusters uh and inspect the hem catalog we'll go to that in a second. Likewise, in satellite, I can create my location on any infrastructure, whether it's aws or on-prem, and in a similar manner. I can actually create an openshift cluster and then deploy the add-on there as well.

C

We'll do a quick uh overview of the openshift cluster, because I built that up when I drive home the point. So if I click on clusters, you can see that I already have a collection of openshift clusters already created and what you can see is uh I'm going to drill down into a ocs cluster name, uh opengl cluster name, stat ocs cloud pack for data id and what you'll see is before I do that. I wanted to kind of give you a the experience of how to create a cluster. Today within ibm cloud.

C

You can select the different versions that you support. Four seven to four five um and you can run on any infrastructure as I've been discussing, uh whether it's classic vpc or satellite, uh enabled infrastructure, uh and you can make various customizations I'm not going to go through all these customizations and by by default we kind of do a three-way zone for application of uh ordering, and so uh you can actually customize that as well.

C

Since I've already created the cluster and that can take you know more than uh up to 10 minutes, uh let's kind of just click on my existing cluster and what you'll see is on my existing cluster. You can see that I have a three zone cluster and uh the health of it is fine and what you'll see in the top level overview page is a bunch of apples over here. One of these add-ons I've already enabled or just to enable an add-on.

C

uh You just click, install I've already done that, and so uh I'm gonna switch to uh the page where I I can actually uh load the openshift console, uh which I've already loaded up for you guys and what you can see is.

C

I've already have ocs installed through the add-on, and in that add-on you can see that there is an application consuming that already, which is uh crop up for data. So if I drill down into it, um what you can see is uh it has quite a lot of pods running and uh a collection of pvcs. If I drill down on the pvcs, what you'll see is all the compactor data pods are running on top of ocs.

C

uh Some are using the file system and some are using the block uh storage for edge use cases. So this is uh basically cloud pack running on uh ocs and if I was drilling down into the cloudtalk application, I had the running application.

C

I've actually fed it with some data with some that analytic data as well- and uh I just want to show you that I do have a notebook created for my data and ai needs uh and trying to show you some analysis here. um So the whole point of this uh part of the demo was to share. You can create and run a mission critical workload, not just a toy application, an ibm cloud with ocs such as a substantial application, such as a profile.

C

So with that context, I'd like to kind of uh transfer this uh part of the demo next part of the demo to akash, thanks for listening to me. Thank you.

D

Yeah thanks uh something, so let me start my screen sharing.

D

So uh here in in this part, I'll be covering mainly around the ibm add-on. So I think sandy mentioned in this explanation that we have created an ibm ibm created ocs operator. So we have deployed this sources operator as an add-on and I'll be focusing more on the use case of why this has been developed and how it uses the verses department.

D

So, as we know today, uh ocs can be deployed even from the web console or even from the cli, and there are different artifacts which have to be created. uh So I'm just in my current screen.

D

You can see some of the list of resources or artifacts on the right hand, side where the ocs deployment has to be done to create them like operator, group, subscription, storage, cluster and a host of resources, and each of these resources need one or multiple parameters which has to be given as inputs by the customers or users, and they have to be created in a particular order, and only once the one resource is successful.

D

We can move on to the next one as as we found out that this, this is having a lot of resources or manifests which has to be handled, and this is uh having some user errors, the possibility of user errors.

D

So in order to handle that we came up with an idea of operator uh which will basically take all the inputs required for those deployment as as inputs for our custom resource and once the customer will deploy the custom resource, we take all the inputs and internally, our operator will deploy the ocs and all the required resources. So we also consolidate the status of all these resources and provide one final status when once the ocs is ready to be consumed.

D

As we can see here on the screen on the left side of the table, we have the parameters which we are exposing for our ibm add-on. So as we see, these are a handful of parameters which are just required to deploy the ocs and we have just done a mapping of uh which uh artifacts on the ocs side, these parameters are going on, so we have some of them like uh the device paths, the mon classes, the osd classes, etc.

D

So this is uh going to ease out the uh user experience and the customers will be only dealing with a single custom custom resource, so in the next one, I'm showing how this will result the number of steps so yeah. This is one of the uh reference of uh installation of forces. On the left hand, side, we have the add-on steps. As we can see. It's only one step here.

D

The customer will create the ocs cluster, which is the custom resource and that will take all the inputs and the ocs will be deployed, uh whereas if you go the cpl iv or the manual way on the right hand, side, we have around eight steps that we listed when we try it out. So this really reduces the number of steps and also the number of resources that the customer has to be aware of to take care of versus deployment.

D

So this is about the ocs install steps, and this is a comparison for upgrading of ocs, so whenever we upgrade ocs from one version to the other version. uh So here also, we have to deal with only one ocs cluster customer source, where we just edit the resource, and we set a boolean parameter as true and once upon, saving that cluster resource. It will automatically get upgraded uh to the next available ocs version.

D

uh Whatever is available so or if you compare it to the manual steps, the customer needs to know which version he has to upgrade and he has to uh provide those things in the storage cluster resource and he has to update the channel so that so those are some of the simplifications we have provided for upgrade steps.

D

So uh this is the similarly uninstall of ocs.

C

D

So, as we can see, uh we have only one step for uninstall of ocs, so uh if the user simply deletes the ocs cluster, that will delete all the resources which were deployed as part of ocs, so it includes storage, cluster set cluster subscriptions and and uh even the local volumes, if there are uh local ones being used in case of the ocs, so that is uh uh easing out uh uninstall procedure or a cleanup procedure in case of any issues for the users.

D

So uh here I'm showing the steps for the scaling of ocs, so in case of scaling. Also, we just deal with the ocs cluster resource where we just increase. The number of osds that are read so number of osd is a parameter where we uh now, where we give the parameter uh customer has to say only how many osd is this. This is mainly about the usable storage that is uh specifying uh so once uh there is, some number of osd is specified or it is increased.

D

uh We we uh increase it by three x number of times, considering the count parameter, which is on the on the right hand, side that you see right so so this is some of the simplifications uh we are providing for the different stages of ocs deployment and uh and upgrade as well. So now, uh as we saw in in this entire exercise, the customer is only aware of one custom resource called ocs cluster right.

D

So the number of steps that we have here is uh there are two types in which we can use this social operator. uh One way is we enable this add-on uh using our ibm add-ons. uh I think that is uh our friendly uh show earlier. I can also show that in the screen, so we have uh different types of uh add-ons here: the openshift container storage, so we have to enable this add-on and after enabling the user has to create the custom resource uh with ocs cluster and all the parameters.

D

So this is one way of deploying the ocs, and recently we provided another option that that will further simplify this deployment of ocs. We call it as a one-click install, so what that means is uh if the user is going to deploy the ocs right after the enablement of the operator, uh he can as well do it along with the operator enablement. So, basically, what that means is uh when the user is going to enable the ocs add-on. He will also provide all the oper parameters required for deployment of ocs.

D

So internally, what happens in the back end is it will first deploy the add-on the operator gets created and then our operator will pick up all those parameters which are supplied and it will declare those cs. So it's it's in a single click. It is creating or deploying ocs, uh but we have also made it as optional, so the users will have two options to deploy ocs uh in this case, so I'll be uh showing a small demo of uh one click install here to start with.

D

So let me show the different options that that we have in the add-on here.

D

So, as we can see, these are all the ocs related parameters we have exposed to the users and in case the user wants to enable the add-on and also deploy the the user can as well use all these parameters, and we also have the default values specified, so the users can take benefit of default values or they can override with their own custom values.

D

So one important parameter I would like to highlight here is the ocs deploy. uh This is one of the parameters we have introduced here. As we see, the default value is false, so what that means is uh when the user enables the add-on and he doesn't give any options, and if this was his deploy is set to false. uh It will only enable the add-on and deploy our operator. It won't deploy the ocs uh so in order to deploy the ocs in the one-click install.

D

This parameter has to be set to true and along with that, the other parameters can be overwritten, so I'll be showing the command that we are using here.

D

So, as we can see uh here, I'm enabling the openshift container storage add-on on my clusters, and here I've set the ocs deploy as as true okay, and then there is those osd size and just for the purpose of demo, I'm taking one parameter of voice to size where I'm overwriting, this 250 gb to 500 gb.

D

Okay. So now, uh when we deploy this, uh if we have to show the add-on which is deployed, uh it will show uh that the add-on is getting deployed and after that the ocs is also getting deployed. So here I'm showing that our uh oc openshift contains storage add-on.

B

D

And we are using the 4.7 ocs version here, so I can show the version of our operator.

D

So this is the operator port, which I am uh where we have deployed this operator, uh and it is actually taking all these input parameters and deploying ocs.

D

So, as we can see here, we are pointing to uh stable forward seven, and that is os 4.7 uh version, and we also see all the other parameters which are here and with this it has deployed the ocs. So I'll show you the ocs cluster that we have so here. This is the ocs cluster I was.

D

uh I was talking about where this is the custom resource that we have exposed to the end users, and this is the single resource which will take all the parameters in our spec as as we saw, these are some of the parameters we gave as at the time of installation, and we gave the site osd sizes 512, which we over uh overwritten or at the time of enablement, and this is the status I was speaking about, where we show one consolidated status for the storage cluster, uh saying it's ready uh and it is deployed.

D

uh So now it is in the consumable state. So I can show uh a highlights of how many different resources are getting deployed uh in case of open shift here. So these are all the different types of uh deployments. Stateful sets demonstrates getting created for mon devices for osg devices, so so these are some of the artifacts I spoke about earlier, which has to be deployed by the user, and one important resource is a storage cluster which is on the ocs side.

D

Here, as we can see, it shows the ready state, which means it is ready to be used, so uh so these are some of the benefits I I wanted to cover in this demo and highlight on the user experience using our ibm ocs operator. I think I will now hand over to sirisha for covering the satellite related demo uh using ocs.

D

Thank you very much.

A

Yeah alright, so let me share my screen.

A

Hi, I'm shirisha and for the next 10 minutes I'll, be walking us through deploying ocs across multiple clusters using the satellite storage templates. um To begin with, this is my satellite location, where I've attached aws instances as hosts and I've also assigned these hosts to two of my ocp clusters.

A

um So these are the two clusters that I've assigned those hosts to, and additionally I've added both these clusters to this cluster group. So we can see that both of these belong here and now I can deploy ocs across all clusters belonging to this group yeah. So the satellite storage templates are provided and tested by ibm or third-party vendors and ibm provides us with the tooling to install these storage drivers on satellite clusters in our satellite locations, based on as described in the template.

A

um If your preferred storage provider does not have any template uh templates present, then we can create our own storage configuration template.

A

um So these are the list of satellite storage templates currently available under ocs we have the ocs local and the ocs remote templates, the ocs local template can be used when we want to use the local storage present on our nodes, and we can use the ocs remote template when we want to use remote block volumes uh in this demo I'll be using the ocs remote template.

A

So the I've already installed aws ebs csi driver on both my clusters using the aw cbs csi driver template. This will allow us to provision remote block volumes. um The first step towards using the satellite storage template would be creating the satellite storage configuration.

A

So when we create a satellite storage configuration, we include the details of a set of parameters that are specific to that template. So this would be the command I'll be using today to create the satellite storage configuration.

A

So here I've provided some of values for some of the parameters like mod size, osd size, one storage, class name and ost storage class name. um So this storage class was created as a part of our aw cbs csi driver install yeah. We can also view this in the ui, so these are my configurations, and this is the configuration that was created yeah, so we can also see the parameters we provided here in the data section.

A

So once we create our storage configuration, we can use this configuration to assign storage across our satellite cluster, so this would be the command I would use to assign the storage configuration.

A

Yeah so here I've provided the group that I've added my clusters to and I've provided the configuration name that we had created earlier and I've given a name for the assignment.

A

So we can also view this in the ui. So here we can see that the subscription has been created and it's been deployed on both the satellite clusters. um Now let us take a look at some of the resources deployed when we assign the storage uh configuration so internally. This the template uses the ibm ocs operator that akash mentioned. So we can look for the ocs operator.

A

A

We can see that the pod has been created and it's running also. We have the custom resource, that's created.

A

And in this text section we can see that it's taken the parameters that we provided while create creating the configuration. Now let us verify ocs installation. So let us look at the pods in the open shift, storage.

A

Namespace we can see that all the pods are running and let me also show the storage.

A

A

Yeah, so we can see that the storage cluster is also in ready state I'll quickly also show the web console for both these clusters.

A

So this was one of the clusters belonging to that group, and if we see the web console, we can see that ocs has been successfully installed and we can get other details in the persistent storage tab. um Similarly, this was the other cluster that was a part of that cluster group, and if we check the web console of that cluster, we can see that again. Ocs has been installed successfully.

A

So with satellite we can bring our own infrastructure to ibm cloud and run ibm cloud services on it. um The satellite storage templates can be used on multiple types of infrastructure like aws, azure, gcp and on-prem devices, amongst others. um The. In conclusion, the advantage of using satellite storage templates is that it provides us with the ability to create a storage configuration that can be deployed across multiple clusters instead of deploying ocs on every single cluster individually.

A

So this concludes the demonstration and now uh I'd like to hand the control back to sandy. Thank you.

C

uh Thank you, um everybody. um I wanted to kind of uh cover the fact that uh uh thank you for akash and sharish for doing the demos and, uh if there's any I'd like to open up for any questions, I see into some healthy discussion on the chat. uh Do you want to ask any more questions? Live we'd, be happy to take them at this point.

B

So sandy I wanted to make sure we addressed something first, because in your slides we didn't mention this at the beginning of the call. So openshift container storage is going through a rebranding to open shift data foundation. So I just want to put that out there as people look at the slides after uh the briefing that if you see odf, that is also ocs. That is.

C

Correct yeah: this is the branding yeah so, and the branding is partly because there's going to be these two, you know uh basic capability and advanced capability, and those advanced capabilities will grow over time uh with additional capabilities as well as the basic capabilities uh is my understanding, uh but the actual operator names are not going to change like what like showed. You know from a technical resource, sampling, okay,.

C

Did I did I address that correctly.

B

I I believe you addressed that correctly. I just wanted to throw that out there, because you know there's both in slides and we'll see that across you know not just these presentations but many different assets. So um so, looking back at the questions, we had some great questions and akash addressed some of them in the chat, but I wanted to do it live as well, so sandy can have some input and for those that aren't viewing this chat, the t-shirt sizes are a combination of cluster size and the number of nodes and we're.

B

Yet to finalize oh see, I'm reading your answer, akash all right. Let's go back to keith's question: can you expand on the ocs t-shirt sizes? What's the minimum and then what's the maximum of the cluster size? Is this in terabytes or nodes go ahead and.

C

Gosh uh I've been uh we've been thinking more about the the size itself and then based on the size. We would calculate how many osds you need, based on the environment, that you're deploying on, so that you don't have to figure out all those choices and all the storage classes, because many.

D

C

Are application users they're, not storage users, and they don't even know what sometimes, what block storage is they just want to read, write or read only interface right for their pvc? That's all they really are.

B

Yeah exactly would you like to add anything akash.

D

uh No, I think.

B

Sandy covered.

D

It so I'm fine.

B

Okay, all right so another great question: um you showed the demo showed a lot of the great features. So can you talk about what dr features so disaster recovery? Do you have you know, do you have replication? You have integration with backups. Can you discuss that further.

C

Okay, so the uh within the essentials package that we'll be offering it does have, you know, uh replication within the zonal across a metro area right because that's uh basically uh we're within a metro cluster. So that's all synchronous, there's work involved with ocs, where they're actually developing uh multi-cluster, dr both asynchronous and synchronous capabilities- uh and you know there will.

C

There will be some time difference between you know, copying the whole cluster state or partial cluster state from one cluster to another uh and that that in that is including the actual data and the application resources. So you can make one cluster uh be equivalent to the other cluster and that could be done across regions right.

C

That is still going to be coming out a little later this year, but that's kind of how we're going to be offering it we'll initially operate the essentials package and then roll out the advanced package.

C

As all those capabilities come up.

B

Thank you, that's good to hear that that's on the roadmap very important features. I know a lot of people are looking for. um Let's roll into there's another question about csi, so many storage providers have csi plugins. Can you share the top two benefits while using ocs this plugin.

C

Yeah so ocs uh I mean a lot of storage providers, they're either hardware systems uh and so they're. Very they have a data gravity in to where they are located.

C

A software defined solution such as ocs, can be deployed just along with your application and as a result, you could take your storage and your application and deploy it anywhere. So you have portability in one of its benefits and then many of the container native storage solutions uh have advanced capabilities that I'm mentioning uh you know such as, like the security features uh that are continue to continue to grow because people care about encryption, they care about you using the native p key management systems that are out there as a deploy of different providers.

C

They care about. You know, disaster recovery. All those capabilities are typically not provided by csi drivers themselves, so ocs is more or odf, as korean essay is like a overall packaging of capabilities. Csi is one of the capabilities that ocs is providing for its storage feature, but it'll be a growing list over time.

B

Thank you. Let's see, you know we had a good question about, so ocs is managed through its operator. How is the csi storage managed you kind of touched a little bit, but.

C

um Hey go ahead. Akash can take that.

D

Yeah so uh in case of uh operator, as I said, uh the operator is kind of a wrapper on top of the existing ocs operator or the css driver, so it is uh making the user experience much easier now releasing the user errors.

D

In the initial days, when we didn't have this operator, we had experience of many user errors and queries uh from our users. So basically, this this operator is simplifying the ocs deployment. So you can consider this as like a wrapper on top of the existing operator.

B

I'm curious about um if you've already figured out, so I'm just looking through the chat about a great chat going on the the billing and that's something that you know a lot of people.

B

Sometimes they get surprised with building, especially when it comes to storage as you're, you know replicating volumes or so how do you have billing set up and how granular is it? um Can one of you talk about billing briefly,.

C

Sure, when we roll out our ga, because right now we're in tech preview phase and we don't actually charge for anything. So it's a great time to try it out and the same experience when we do billing it'll be based on the nodes where you have ocs installed or odf and will be based on the types of nodes, whether it's a virtual instance of vsi and will be calculated based on four cpus per vsi.

C

We're still coming up with the actual charge for uh that that metric. But once we have that uh billing charges will be calculated based on that and then for bare metal uh it'll be a straight charge. Monthly as well, and it'll be based on node instances. So wherever you install the ocs is, is really the number of nodes and there'll be a charge for that, and you can use it as long as you want and tear it down, and you only get to pay for what you need. Initially.

C

You'll have uh hourly plans for vms and then monthly plans or bare metals.

C

We may look at future other plans in the future, but at this point those are the current plans that we're targeting at least initially and with billing. You get the support right, both from ibm and red hat.

B

um But so support through ibm and red hat, you first go to ibm to make sure.

C

B

Purchasing through ibm cloud, correct.

C

And you can use ibm ticketing system to open tickets, but you know, um but we're backed by red hat as well, uh and teams there as well.

B

Now sharisha, I didn't have a question to ask: you: is there something that you would like to bring up in our remaining minutes?

B

Put you on the spot.

A

I don't know I don't have anything to think of right now. Everything's been answered.

B

I don't want to have you feel left out all right, akash sandy is there anything that you would like to leave the audience with some key talking points, key business objectives on using ocs with ibm cloud.

C

um It I would say, if you're looking for a solution, that's already integrated and easy to deploy. You should consider that we do have other alternatives in ibm cloud. You can look at those, but with ocs, it's integrated into the user experience and we've simplified the deployment across those kind of environments where we run it and if you're thinking about a hybrid environment, ocs is a good choice with ibm offerings using red hat, so give it a try and give us some feedback.

B

And thanks for walking us through the technical being able to deploy it and scale it. That's all really important for all of us to see. So. Thank you all right. Thanks.

C

B

Not any more questions then we'll have chris see us out on the live streams. Thank you all.

B