Parallels RAS foundations: Mastering templates, clones, and providers

out of four technical webinars that we are doing around Parallels RAS technical professional trainings. As same as previous two sessions, again a very packed agenda. Today everything is about mastering templates, clones and providers. So we have basically four chapters we want to go through. Templates, clones and providers, that's I think the most important one for today. Then we're going to look specifically also at some things which are a little bit different for Azure Virtual Desktops or some things that we have added specifically for Azure Virtual Desktop. Then we're going to continue with some troubleshooting because we noticed that if you take a look at support and where tickets are open, it's very often for templates and clones. So we're going to help you with a bit of troubleshooting if that would be needed. And then a very short end around backups, upgrades and account management to wrap up today's session. Again, lots of people live today. If there would be any questions that you have, there is a Q&A possibility in the webinar. So don't hesitate to ask questions and we'll try to answer them as good as possible during this webinar. With that said, let's start. So what we're going to do today is mainly a mix of going through some slides. But previously, like last week, I'm going to do a lot of things also live on the system. So let's start with a small recap of what providers are and what they can do. And then we'll start deploying some of them and go into the details. So last time we created mainly a static environment. If you still remember from last week, we created two session hosts, we created two gateways, we created two brokers, but that was like a static environment. What these providers can do is make you a dynamic environment. So these providers or connectors or agents that integrate with underlying supported hypervisors or cloud platforms, they do a few things. One of the things they can do is provisioning. We can, for example, say create me an extra VDI, create me an extra session host. They can also do things like power control, for example, do a reboot of the servers at four o'clock in the night. And they can also be part of things like auto scaling, for which we can then do things like cloning and working with the templates. And also something we have mentioned that I'm going to show you straight away is that these provider agents typically would run from the connection broker. So typically you would not need to install any other virtual machine or another piece of infrastructure. Typically they would be the built-in versions that we're going to use. And that's the same I'm going to do today. So we're going to use the built-in provider agents on the connection brokers, but it's also possible to do dedicated deployments. And that's also something I'm going to show you later on. Another thing I want to highlight. So today we're mainly going to talk about what we call tier one integrations via these dedicated providers. So this webinar was recorded when version 21.2 beta was available. So the webinar is mainly talking about the capabilities at this point. And if you take a look, we have eight providers available today. So we have on one hand VMware, on the other hand Hyper-V. Then we have Microsoft Azure and Microsoft Azure AD. So it's actually two different ones. Then we have one for AWS. And then this is something which is new in ROS 21.2. So we used to have already a provider for Nutanix AOS, so which basically integrates into Prism Element. But Nutanix is moving all their capabilities and all their functionality to their Prism Central platform. They have also like a new set of APIs and they're going to deprecate the old APIs later this year. So as of this 21.2 release, we have also done like no native integration in Prism Central. We have also integrations with scale computing. And then there's a last one, which is a sort of remote PC provider, which allows us, for example, to do things like giving people access to physical machines or to a pool machine. So that's also like a little bit of a special case. So this is what we are going to cover today in this webinar. Something I'm not going to cover, which is also new in ROS 21.2, is the fact that we now have an open provider framework. So basically what we now can do with, since ROS 21.2 is integrated to any hypervisor. So prior to ROS 21.2, you could already install ROS on any hypervisor. That was not a problem, but we didn't have like a mean to integrate with the hypervisor or the underlying flow to do things like power management or templates, cloning, scaling, these kinds of things were not possible out of ROS itself. You could do that, of course, with some scripts and some other tools out of ROS, but this is something we have added in ROS 21.2 that we have now like an open provider framework that we can integrate basically in any cloud or any hypervisor. So this is something that will be covered in a dedicated webinar that we're going to announce in the next weeks. I'm just going to touch it very briefly in this webinar to already do a little bit of a teasing, but the main webinar on the details will be later. So maybe before I go into the slides, let me just show it live, how it looks like, and then we can use the slides as a sort of resume. So if you want to integrate into underlying hypervisors or clouds, you have to go to your management console, and there is a dedicated tab called provider. And if you click on provider, and then on plus, you can choose whatever hypervisor or cloud platform you want to integrate. So for example, if I want to integrate with the unit NX Prism Central, I have to provide some details. As you can see, I have to give a name, I have to give the address of my console, I have to give a username and password, and then basically I can click on next and the integration would be done. Same for example, if you would integrate into like a cloud platform, and let's for example say I want to integrate into Amazon, then it will not ask for a username and a password, but it will ask for an access key and a secret. But this is a little bit specific for every hypervisor or cloud that we want to integrate. You have to give the credentials or the way to connect, and then once that is done, the provider will be automatically connecting. So for example, in my environment, what I have done is I have for example set up an environment which is like integrated into Azure Virtual Desktop. The custom provider is a little bit different. So if I would like to deploy one of these custom providers, you will see that it's different in the sense that what it's going to ask is a few settings. So it's going to ask like the username and the password as which this script will run, and then you will have to provide also some links to the script itself. So I have to provide credentials, and then on the connection, I don't have to specify like an IP address or anything like that, but I have to specify the scripts that can be used and to integrate to these APIs of these other platforms. So in this case, it's a bunch of PowerShell scripts that I'm going to use. A bit more about this in another webinar. So once this is done and everything is green, it means that as of now, we can start creating a dynamic environment. So let me switch back to the slides. So let's go back to the slides. So for deploying a provider, what we need to do is, as you can see, so first of all, select the one you want to install, then based on the type you want to deploy, you will be asked for a username or password, or you will be asked for more cloud-specific connections like an application ID and a secret key. Something I didn't mention, under advanced settings, if you click on advanced settings, it will give you the opportunity to pick either the connection broker you want to dedicate to this provider, or it will also allow you to use a dedicated provider agent, and then you can specify another box where this provider will be deployed for. And then once this is finished, you will see it will pop up. So very simple for these T1 provider installations. For the open provider framework, so that's the thing that we're going to cover in a dedicated webinar. It's similar, so you provide the credentials for which user these scripts will be run. You provide the links to the scripts. Parallels will provide some sample scripts. One of the sample scripts we will provide is for Proxmox. We will also provide some scripts for testing and debugging. So if you want, for example, to develop your own set of scripts with Google Cloud, with KVM, with any other platform, we will also provide you some test scripts for doing that. The next step, this is something I didn't show you live, but then once these scripts are set up, you can pass variables and values. So you could, for example, say this is my username, this is my password, this is the host I want to connect to. So if you have multiple Proxmox clusters, you could then, for example, pass the IP address of the console or these kind of things, and then just deploy. So basically, it's almost the same, except that in this case, you create your own integration. Once we have this provider set up, then the next thing we need to do is create a template. So to create a template, let me just go back here and do it also live. So let's, for example, if I want to create a template, I can create templates both for session hosts and for VDIs. But let's say, for example, I want to create a template for a session host. I can click on plus, and because we have this provider integration, it will, first of all, list all the machines that are available. So let's, for example, say that I want to make this Ruslop02 my template. Then let's click on next. I can give it a name. Then, of course, based on the underlying hypervisor or code, it will give me a few options, like this is now running on Azure, so I can only do full clones. But if you would be using Nutanix or VMware, it would, for example, give you the opportunity to go for a linked clone instead of a full clone. It will ask me some details, like, for example, how big does this... So if you make a clone from this template, which size does it need to be? As this is, again, Azure, it's asking me Azure-specific sizing. If this would be VMware, it will ask you, like, how many CPUs, how much memory, which storage group do I want to use to deploy these clones. So it will, of course, change the parameters depending on where you want to deploy. Then it's going to ask me if I want to rasprep or csprep this environment. I'm going to explain later the differences, but this is basically initializing the template for copies. I need to specify, like, the local account that's going to be created whenever I create them, and then, of course, also an account that we can add them into the domain. And then from here, I'm going to stop here on the live machine, but then there are a few more steps that need to be done. But it's a very simple process. So you just take, like, one existing machine on your hypervisor or clone that is visible with that provider, and you turn that machine into a template. So let me just switch back here to the slides this time. So this is something you already have seen. So I have my machines. I give it a name. I give it the correct size. I choose rasprep or csprep. I give it all the details. And then once we are through this step, this is something you already have seen last week. And when we were doing the webinar last week, we were able to decide which optimizations we wanted to enable, so we can do that also on the template. We can select how to do the license management on the Microsoft site if you want to use, like, a key management service or a Mac service. And then the last step before we deploy is to give it, to give, like, the version a name. So whenever you create a template, you create the first version of that template. So we have called it this time the base version, and it has, like, version ID number one. We can also give it, like, a tag. We could, for example, say, like, this version of the template is the production version. This is something optional, but you could, for example, also tag it as production. And then if you click on finish, it will create the template. So this is something I have done for VDI. With this template created, the next thing we can then do is create a dynamic host pool. I said I would come back on the differences between Rasprep and Sysprep. So I think most of you are familiar with Sysprep. So Sysprep is the Microsoft tool for turning a machine into an image, and then from that image, you could create, like, new copies of that machine. Rasprep and Sysprep do more or less the same. There are a few subtle differences. I think the most important one is that Rasprep is a little bit faster than Sysprep. I think Sysprep typically needs to reboot at least twice. Rasprep only reboots once, and there are a few things which are not done. For example, if it's, like, in your case, important to have, like, a unique SID, then I would say, like, use Sysprep. If that's not important, then Rasprep would typically do the job. But they are both processes to make an existing machine into, like, a machine that can be used for copying. So once that template is done, we can create a pool. So let me also switch back and do this live. So if I go to my VDIs, for example, in this case, you will see that I have already here a template, which is for, like, my Windows 11 template. So if I go here to VDI and then to host pools, I can click on plus. I can call this my – let's call this the live pool. Click on next. It's going to be created from a template. Next. Which template? My Windows 11. Which version? Let's say I want to have the base version. Click on next. It's going to ask me, like, how many VDIs could be deployed. So let's, for example, say that I want to have, like, maximum 10 VDIs. And it's going to ask me, like, how many of these VDIs do you want to be created immediately, like now after completion of the host pool. So let's say, for example, that I want to have, like, two of these machines. And then it's going to ask me, like, what are the names of these machines. So as you can see, for example, this is, like, live pool 001, 002, and so on. So click on next. This is also something which is new in RAS 21.2. You have now the possibility to overwrite the template settings on certain aspects. So, for example, when I was creating my template, I had to provide, like, the details for my active directory for joining that machine into AD. And I had also to specify, like, an OU. Maybe you have, like, different OUs for your production machines and for your test machines. And on these different OUs, you have maybe, like, different GPOs attached to it. So now with one template, you could, for example, deploy in two different OUs and as such have two different GPOs attached to it. Or maybe if you deploy, like, first a test batch, if you're not worrying about, like, high availability. So you don't care if it's, like, in different availability settings because it's your best pool. But maybe this one is, like, your production pool and you want to have, like, on an availability zone. Then these are also things you can overrule. And then the other thing you can overrule is the host size. So prior to 21.2, whenever you created a template, it always had fixed the same host size. As of 21.2, we have capabilities of, for example, deploying in other hosts in different sizing per host. So this is also something you can override if needed. Let's click on next. I can also specify how many machines need to be available in the buffer. So what does that mean? For example, specifically with VDI, if somebody connects and all the VDIs are in use, you typically don't want them to wait before, like, a machine is created. You always want to have, like, some machines available. So what I could, for example, say is, like, I want to have at least always, like, one machine available in the buffer. And then RAS will always make sure that there's, like, at least one machine that is not assigned. And if there is, like, no free machine, it will automatically start cloning already one extra. Once the machines are created, let's keep them powered on. And what do we want to do with machines that are not used? We could, for example, say, like, after 30 minutes, you can delete. So we can set all these settings. This is also something interesting. With Parallels RAS, we can both create, let's call it, persistent desktops and non-persistent desktops. What's the difference? A non-persistent desktop will always reset back to its original state. So what I could, for example, say in here is that, for example, whenever somebody logs off, then recreate the image, recreate the VDI immediately. And what will happen is that on the moment somebody logs off, the VDI will be recreated. So you will have a sort of, like, non-persistent desktop. I can also deploy FSLogix. That's something we have covered in the first webinar in this series. So I'm just going to leave it off. But if I now click on finish and on apply, you will see that I'm going to create, like, a second pool called the LivePool from this template. Good. You will also see if I go here to hosts and I do, like, a refresh, you see it's creating LivePool001, LivePool002. So let's go back to my slides and continue from there. And so deploying a host pool, you give it the name, you specify the amount of machines, how many machines need to stay available, always in the buffer. If you want to overwrite things, that overwrite is something which is new in VDI1.2. And then the final step is then to define if you want to do, like, anything after somebody logs off or disconnects, selecting the profiles. That could be, for example, FSLogix and deploy it. So this is what we have just done. If you want to do the same with a template for a Windows Server version, let's say a multi-session host, then all the steps are the same except the question, for example, like how many machines do you want to keep, like, on standby? That works a little bit different. So if you deploy, like, multi-session hosts, you have to specify also how many machines that need to be deployed once the install is completed. So I could, for example, say, like, deploy five Windows Server 2025. And then there is an option to enable auto-scaling. So if you enable auto-scaling, you will have a bunch of options. And then one of the options is, for example, to specify that the minimum number of hosts to be added. So you could, for example, say, like, whenever, in this case, for example, whenever the load hits 75%, then add one machine. How many machines need to stay all the time? At least one, for example. How many can there be maximum in the pool? Like, 10. And whenever you hit, like, 20% load, at that point in time, you can start training machines and you can get machines from the pool. And then, for example, once the machines are removed from the pool and drained and powered off, you can, for example, enable that. So this is, for example, one way to save costs on Azure. You could, for example, have during the night only, like, one host pool, one machine in your host pool available. And then during, for example, during the day when the load is high, you can add more machines, as many as you need for having, like, environments which are, like, so templates is more or less the same on the VDI or under RDSH. It's just like this auto-scanning that is a little bit. And once we have done this, then, of course, we can start publishing resources. So let me also do this live. So if I go to publishing, I can click on add, I can choose a desktop. This time, it's not an RDSH session host desktop that I'm going to deploy, but I'm going to deploy a virtual desktop from, for example, my VDI pool. I'm going to deploy it. And then I can call this, like, my Windows 11. I'm going to specify, for example, all the available size. And this is also, like, a little detail, enable static assignment to host pool. That means that whenever a user logs in, we will always assign the same machine back to him. But this is, for example, where you can specify if you want to use with static pools or with dynamic pools. If you switch this from enable static assignment off, then whenever somebody logs in, they will get, like, a random machine. If you have it on, then it will be always the same machine that is assigned. If I click on finish and on apply, then I have my desktop available. So, for example, let's have a quick look. So if I go here to the farm and to VDI, host pool, my pool, what I have specified, I think it's here under my settings. You see, it says always, like, keep one available in the buffer. And I have for the moment only one, I think, which is there. Yeah, there's only one. So what will happen is that, let's say, I log in with lab one user. So let's log in. No, it's still disabled. So I have my Windows 11 desktop. If I click on it, yeah, I'm going to open it. It's okay. Make no difference. Yeah, one second. Let me just start it this way. Start it this way. Okay, yeah, that's where my live demo goes wrong. But let me give it one more. I think it was, like, also, let's switch the slides. I think it has to do with the fact that my user is not authorized. Let's switch back to the slides. I'll show you on the slides. So you can publish. And then once it's published, what we can do is this. Let me fix it another way. Let me just use another user. So when I start this, this should normally work. So I'm connecting. My desktop should be there pretty soon. Let's go back. Let's have a look what's going on. So looks like, okay, this is the thing with live demos. It looks like my lab one user was assigned to that first one. So there was no other machine available. So my lab two user is now assigned to a second one. And in the meanwhile, he's creating, like, a third one. So he's cloning this one because he always wants to keep one available in the buffer. And I think if everything goes well, you see, I'm logging in. It will take a little bit of time because it's the first time I log in with this user. So it will create a profile and these kind of things. But basically, you see everything in action. You see, like, my lab one user is assigned to this one. My lab two user is assigned to that one. And because there was, like, no other machine anymore in the pool, it's automatically creating one. So as you can see, very, very dynamic that these machines are spin up and started. Let's switch back to the slides and show you a few other things. So yeah, the publishing we have done. Something we also can do with RAS is add things to the scheduler. So the scheduler is here. We can create actions based on individual hosts, or we can also create actions on the pool. So, for example, maybe you want to reboot all your machines on Sunday night. Then you can, for example, say here, like, reboot host pool. Which pool? For example, my live pool. And then I can create a trigger. So that trigger could be, for example, every Sunday. So let's take the 24th. Maybe I want to do this at 2.30 a.m. And we need to specify how long it can take. Let's say, for example, this needs to finish in one hour. And we want to do this every week. And then we can also add some triggers. So, for example, one of the things we could do is enable a message. So servers are going to reboot for this weekly preset, for example. So we're going to set this 15 minutes up front. And then what we can also do here is, like, enable drain mode. So we could, for example, say, like, if somebody's connected to the system at 2.30 in the morning, we're not going to kick them off, but we're going to give them, like, 30 minutes to log off. And then we're going to do, like, a forced reboot anyway. So with the scheduler, we can do these kinds of things. Let's switch back to the slides. Okay. So the scheduler is there. Good. The other thing I need to show you is versions. So we can create per template different versions. To create a new version of a machine, you have to put the template in maintenance mode. So if you do a right-click on it, you can put it in maintenance mode. As soon as it's in maintenance mode, you can make the changes. And then if you click again on it, it will go out of maintenance mode, and it will ask you if it needs to create a new template. So this is what you can see here on the slides. So you click on bring it in maintenance mode. Making the changes is basically connecting to the machine, maybe installing, like, a new software or updating, like, some patches. And then if you click again on maintenance, it will bring it out of maintenance. And what it will do is then ask you, like, do you want to create a new version of the template? So then you have, like, two options. You can create a new version, or you can discard your changes. If you create a new version, it will ask you to give it a name. You will be, again, able to create a tag to it. And then finally, you can click on versions, and it will show you, like, all the versions of this template. Versions can be attached to each other. So you could, for example, say, like, version two is created from version one, version three is created from version two. Maybe in your version three, there is, like, an error, and you need to start over. So then you can, for example, say, like, when you put it in maintenance mode, it will ask you, like, do you want to put, like, version one in maintenance mode, or do you want to put version two in maintenance mode as a starting point? Then you could, for example, say, like, I want to put, like, version two into maintenance mode because I want to create, like, a version four that was created from version two and not from version three. It also supports up to five versions. So if you go above that, you will need to delete one of the versions. All the things I've shown you work for all the tier one providers. There are a few specifics I want to cover for AVD. On AVD, we do a few things a little bit different. So with AVD, we, of course, do integration into the AVD broker, but also on the client side, we do a few things differently. So the Parallels RAS client will, in case of AVD resources, detect that it's an AVD resource, that it's a resource that is provided over AVD, and it will make the connection over the AVD connection broker. So if it's, for example, something that runs on native Azure or on VMware or on anything in your data center, it will use the gateways from RAS. But, for example, if it's an AVD resource, it will figure out what is the most appropriate way for making the connection, and it could be that it's, for example, sending things over the AVD connection broker. What do you need to know, or what are the prerequisites before we can start with this? So the initial things you need to know are that... Sorry for this. So the initial things you need to know is that, of course, you need an Azure Virtual Desktop license entitlement, that's minimum. You also need to configure Microsoft's Entra ID, and it's also highly recommended to configure the single sign-on and also the synchronization between your Entra ID and your local AD. So these are all prerequisites that need to be done. The Parallels Windows client is compatible with the Microsoft AVD client. So basically, the Microsoft AVD client or the Windows client, as it's called these days, is part of the Parallels client. So if you connect to an AVD resource, Parallels will automatically use the Windows client. If you would connect to an AVD resource, for example, on Linux or on Mac, and you're using, for example, the native client for Mac for Parallels, if it's an AVD resource, we will then automatically, for example, switch to the web client for those kinds of resources and do the integration that way. So all of these things are done transparent. But there's also something you need to know, and you might have seen that it was not visible in the list of providers. So if you want to make the AVD visible, you have to go here to Features and then click here on Enable Azure Virtual Desktop. This is something client-related. This is something that was important prior to RAS 21.2, because some of the features were only in the Parallels client and not in the Mixed client with AVD. But this is like deprecated. So what I would highly recommend, and it's I think also stated here, keep it on standard. There's no real reason anymore to use the advanced client settings. As soon as you do that, you will see that over here, the Azure Virtual Desktop becomes visible. And then you can, from here, then also do things like templates, host pools, workspaces. So it's completely customized to not only manage, but also automate AVD deployments. But for the rest, it's almost the same. There are a few other things, I think, which are worth mentioning when it comes to AVD. So with AVD, we can do, of course, cost optimization. So we can do autoscaling, power management. That's very similar, like I have demonstrated on any other hypervisor. Something which is specific for AVD and Azure is that we can also do disk storage cost optimization. So there is a setting under the provider. Let me also show you that live. So if you go to providers, and for example, this one, under advanced, there is a feature that you can enable. And what this feature will do is that, for example, if you have a machine and you shut it down because you don't need it, and maybe the load is below 20%, it still will cost a lot of money on the storage side. So if you have this feature enabled, the storage will be automatically moved to cheaper storage, to HDD storage. And then whenever you need it back, it will be moved back to SSD storage and start it up. So this is something specific for AVD and Azure that we don't have on any of the other providers at this point. So the cost optimization, it's both CPU, memory, and storage. And this is how it looks like. So if you see, for example, this machine is currently running. It's powered on, and you see it's on a premium disk. Then on the moment it's powered off, it's exactly the same disk. You see it is moved to a standard LRS optimized for cost storage disk. And then whenever you need it back, it will be pushed back. Something else which is also available, this is already there since RAS 21. It's a cost dashboard. So this is something reporting and monitoring is something we're going to cover in the last webinar of next week. But for AVD specific, we have also a cost dashboard where we show you in real money where the cost savings of having these auto-scaling and storage optimizations on in RAS versus having, for example, modding them and keeping it running all the time. We show you details as the cost of each of the pools, the compute cost, the storage cost, the bandwidth cost. I see that there is a typo in my slides. It should be, of course, bandwidth cost. And then the savings, like how much was saved on storage and on CPU, thanks to having these settings in parallels. And then we do also like a sort of calculation where we do like a total cost per user, which is like the host pool cost divided by the maximum number of active users. We're going to cover this next week, but for the Azure Virtual Desktop cost dashboard specific, it's part of the RAS installer. So it's not that you have to do anything special. It's just part of the RAS installer. And then once your reporting service is enabled, you can then also enable the cost inside. So this is how you enable it. But these are some very specific things for AVD. As you can see, sometimes things go wrong, even in a live webinar, as I had with that one user that did a little bit strange when I wanted to show you the auto-scanning. So that's why I think it's also important that I spend a little bit of time on troubleshooting. So let's have a quick look at some of the common things that can go wrong and how to fix them. So the first thing which is important if you deploy providers is that the status is OK. So if something is not working as expected, first thing, go to your provider stop check if it's OK. There are different statuses. So of course, if it's green, then it's good. If it says like orange needs update, then it means that the agent of your provider is not in line with the agent of your connection broker. So then it's pretty simple. You do a right mouse click and say update agent. If it's not verified, then it means that there is something really wrong. And if it really states connection failed, then very often it's something like a problem with the network or maybe the username or password. So this is like typical things to look. So for example, if you want to check your agent, you can click on the provider, click on troubleshoot, check agent. And then if it's missing, for example, you can deploy it or if it needs to be upgraded, you can upgrade it from. So this is like the first thing to check. The next thing to check then, of course, is, for example, things like credentials. Maybe you have an account with, maybe you have like an account where the password changes, needs to be changed every 30 days or every 60 days. Of course, you change it on the Windows site, but then you also should not forget to change it on the Parallels Glass site that the account that is used to connect to your provider also needs to be updated. We have a list of supported hypervisors per provider. So it's a sort of matrix where we say like RAS 21.2 has been tested with these providers, with these versions of the hypervisor or the cloud. RAS 19 has been tested with these, RAS 20 has been tested. So there is a list of supported combinations. And then, of course, as mentioned last week with the troubleshooting, there's also the firewall rules. So the ports that need to be open for VMware are different than, for example, the ports that need to be open for Azure. So this is also something to double check. Then once the machine, once the provider is green, then the next part where it can go wrong is on the cloning itself. So sometimes it will say like cloning failed. Then the most easy way to figure out what went wrong is to double click on the host that is more detailed. You will have like some indications where it can go wrong. There are two big, let's call it buckets of problems or issues where things can go wrong. The first one could be really on the cloning itself. Maybe there is not enough disk space or maybe your template is still in maintenance mode. If your template is in maintenance mode, it will not clone or there is something wrong with the sysprep. So maybe it's sysprep that did something wrong. So this is all things on the cloning side. It could be that your clone is created correct. So if you check your hypervisor, you see a machine, you can connect to the machine, everything looks fine. But that Parlas RAS still thinks that there is something wrong with the cloning. That could be that once the machine is up and running, that there is an issue with the connectivity between the Parlas RAS connection broker and the agent that runs on the new liquid machine. So for that, again, it's typically things like firewalls that you need to check. There is also a knowledge base article. That's the last one that I put on the slide. There's also a knowledge base article that shows you how to configure, for example, pools that are in other subnets than the connection broker. So maybe you have your host pools in one subnet, your connection broker in another subnet. Then there are some settings that will help to make sure that the agent finds back the connection broker. Because typically the agent will do a sort of broadcast to find the connection broker. If that broadcast is not working, then there are some settings that we could do to initialize these agents up front to know where their connection broker can be. I think I already covered this for a big part. So use RASprep if it needs to go fast. Use sysprep if things like unique SIDs are important. Or if you, for example, deploy across diverse environments, then it might also be a good idea to use sysprep. If something goes wrong with sysprep, that's of course a Microsoft tool, then the log files can be found in this folder. I've also put here the link to the Microsoft article for more details. For RASprep, make sure that the account that connects to the clone and to the machine has the necessary privileges to do these kinds of actions. So make sure it's an admin account with the needed permissions. Things that also could go wrong from time to time are DNS. So maybe there are still some stale DNS records or maybe you have done like, instead of deleting things in parallels, you have deleted it on the hypervisor and then there might be like a discrepancy between the parallels database and the database in DNS and on the hypervisor. So it could be that you have to do some cleaning up. And then another thing to double check is that, for example, the account that you used to do the domain join, that needs to be a UPN format. So this is also something that people maybe sometimes do like administrator instead of administrator.domain.local. So this is another thing to double check. The last thing on the troubleshooting is that if things go wrong, as usual and as always, the parallels logs are your best friend. So if you take a look on the broker side in the parallels logs folder, there is something called the VDI agent log. And so that's like your first indication where to look if something goes wrong. If you don't find an obvious reason, then maybe again increase the log level from standard to detailed to get more information. The other place to look would be in the connection broker log. So there could also be like indicators where something could go wrong. On the host side, there's also something called a guest agent log. So on the machine that is cloned on that clone, there is also like a guest agent log that will contain details. So maybe make like a combination of looking on one side, looking on the other side to understand what goes wrong and to have an idea where the templates or the clones could fail. I have one last small topic to cover, and I have seen that we had already some questions that came in. So I'm going to cover the questions in a few minutes. So the last little topic for today is about backups, upgrades and account management. So I think the first thing you need to know is that it's of course not needed to give every administrator an account that can do everything. Of course, my demos that I have done today and over the past days were always with an account with root administrator rights, so that that account has full access to everything. But it's also possible to create dedication of rights. So let me maybe also show that live. If you go here to the settings, administration accounts, you can click on plus. You will see that it's possible to create different types of accounts. You can create a root administrator account so that user has all full admin rights. Then there is a power administrator. So the power administrator is a account with a limited set of permissions. He can do things like shadowing sessions, restarting sessions, making small modifications. But that's typically an account that cannot add for example gateways or connection brokers or these kinds of more advanced things. It's an account for limited admin rights. And then of course, especially in a larger organization, it would make sense to make a custom profile. So for example, this is what I have done here. So you can create an account with custom administrator privileges and then you can go per site and also on the global level make specific permissions. So I could, for example, say that this user can do everything on the host pools, but nothing on the other things. Or for example, it can view the gateway changes, but it cannot do anything different. So we can give different privileges per user and per setting. Or on the global side, maybe you want that this user can only do things like reporting and you can activate these settings. So it's very powerful, very granular to set. We can even create a copy. You can assign it to not only specific users, but to certain security groups. So it's something very powerful that can be further fine-tuned. And the last thing, and I'm going over the questions, it's about backups and updates. So if there would be a need to create a backup of your environment, at least not the environment itself, but the settings of your environment, you can click here on export settings and then it will create like a backup file. So let's, for example, say this is my 2005, 26, 16, 50 backup. So if I click on save, it's now going to export all my settings into that file. If there would be like, for one reason that I would need like a reset of all my settings, I can go here to import settings and then I can take my backup file and reset all the settings to the point I've taken a backup. For upgrading RAS itself, so there is a setting here like a check for updates when launching the RAS console. That will tell the administrator whenever there is a new version available. Upgrading RAS is pretty straightforward. So either you get the new version via this check, or you go to the myparallels.com account and you download the new version. And you just run the installer on top of the primary connection broker. So once that is done, your primary connection broker will be updated to the next version. And then as soon as you log in, the Parallels connection broker will tell you like, hey, you have updated the connection broker to a new version. Do you want me to upgrade the other components? And then you can, from there, start upgrading the other components. And we already have covered in one of the advanced webinars how to create, for example, certain architectures that can be upgraded with as less as possible downtime for users. So this is something that we already have covered in one of these more advanced webinars. So this is the last part for today's webinar. I keep opening the slide for the next one. So apologies for that. So what are we going to see next time? So next time, we're going to wrap up this series. And so next time, we're going to do the fourth and last webinar in this four series of webinars. So we're going to talk about things like client policies, universal printing and scanning, session management. So like a bunch of operational things. How can I make sure that, for example, if somebody from sales connects that they can print, if somebody from engineering connects that they cannot print, or how can we configure that you can only connect with the Windows client or only with the HTML client or these kinds of things. So that has everything to do with these client device policies. And we're also going into the concept of teams and then using teams to, for example, create different authentication mechanisms. Maybe you want to have like a team with a single sign-on and you want to have another team with MFA. So this is also something we're going to cover next time. We're also going to cover the HLB load balancer, how that one works, how to configure that one, and how to do a configuration of load balancing on resources. We have like different mechanisms that we can do load balancing towards resources. And then we're going to cover things like reporting, monitoring, and notifications. And then finally, we're going to do like a quick review of the Citrix components and how they map to Parallels components and how typical migration can be done. With that said, I'm through the content, but I see there are still a lot of questions and we still have like a few minutes. So let me switch to the questions and see if I can answer as many as possible still during this live webinar. So let me see. So there's a question on a rollback for upgrades. So that's indeed something like the upgrade is indeed installing over an existing version. I think if you want to do things like a rollback, then probably the most simple way to do that would be to take like a snapshot of your connection broker and your session host and then roll back to that snapshot. I think that would be the most easy way. Then if you have to do like a rollback after an upgrade. So there's a question on Teams optimization. So Teams optimization is supported if Microsoft supports it. So for example, if we're connecting to AVD, then automatically Parallels will also benefit from the same Teams optimization as Microsoft have done on AVD. So it basically depends on what Microsoft has made available. We don't have an optimization pack for Teams like Citrix has, to be very clear. But for example, all the optimization that Microsoft has done in AVD will also be automatically will be available in Parallels RAS. So that's not a problem. I also want to highlight that the Microsoft RDP protocol has enhanced a lot the last year. So we see many users using Teams on Parallels, even in on-site installations without having a lot of problems or a lot of performance issues. There's a question of can you have multiple providers in RAS? Yes, the answer is yes. So the question was very specific on like one to an old VMware version and one to a new one. That's absolutely no problem. Or you can also mix. For example, you can have one to Azure and one to VMware or one to Hyper-V and one to another Hyper-V. Or maybe you have like a VMware cluster in one data center and another VMware cluster completely isolated in another environment. So yes, that's all possible. So there is a question on the release date for 21.2. I'm not a product manager, so this one comes with a little disclaimer, but the plan is to release by end of June. And I think the betas should either be available already today or very shortly. So it should be there very soon. The betas and the final release will be, let's say, second half of June normally. There is a question. Let me read this a bit longer. I think there is a question on something very specific, but I think that's probably something that's best looked at with support. Let me skip that one for now. And then there's a question that I don't know the answer on. The question is, do we have only full clones on AWS or also linked clones? I don't know the answer on that. I would need to look it up. But if AWS supports linked clones, I don't know if that's technically possible, but if it would support it, it might be that we also support it, but I'm not sure about that one. So apologies for not knowing the answer on that last question. But I think I have covered everything which I could find in the chat. So with that said, and I think with being again almost on the hour, thank you for joining me today again. Hopefully see you in the next webinar next week for the final set of this standard TP-Rouse webinar. Enjoy your day. Bye.