Performing a Lift and Shift Migration with AWS Application Migration Service - AWS Virtual Workshop

Key Takeaways

[Music] hi and thank you for joining us today to talk about application migration service also known as mgn i'm going to pause here for a second and already tell you mgn is short for migration the question keeps coming up application migration service or ams is already taken in amazon it's uh amazon managed services so we are mgn um really quickly what we'll be talking about today i'll present myself and my co-presenter uh we'll do an overview of the service we'll also do a detailed migration walkthrough and we'll leave time for questions and answers and we'll touch on some additional resources so uh my name is iron i'm a senior product manager at cloudendure and i'm the owner of mgn and with us today is also pritam who is a senior cloud support engineer with aws premium support and he's a cloud indoor nmg and specialist and for those of you who don't know cloud and door i'll touch on that in a minute so what is application migration service it's the primary migration service recommended by aws for all lift and shift or re-hosting migrations into aws it's highly automated uh so uh the very little uh manual work if at all that you need to do to get your external server working as an ec2 instance you can migrate from any source physical virtual cloud-based servers and you don't have to deal with compatibility issues of making a server that's running on a different type of hardware work on the virtualized hardware which is ec2 instances and all this happens without interrupting your source and uh allowing you to perform cutovers quickly and with uh certainty that it'll go smoothly um mgn makes migrations simpler and therefore it makes them less expensive both in terms of money and in terms of the resources that it takes to migrate to your servers and your applications to aws and it can be used as i said to migrate external servers into aws into ec2 instances running in aws but you can also use it to migrate between two aws regions or even to migrate servers ec2 instances from one aws account to another and it is also the next generation of cloud migration so i know some of you are here today know what platinum migration is but for those who don't filementor was acquired over two years ago by aws it has a migration and a disaster recovery solution that have been in the market for over seven years uh on the migration end we've migrated tens of thousands if not more servers into aws so this is field uh proven technology and let's talk a bit about how it works so you start by installing an agent um the aws replication agent on your source servers and this can be done manually or unattended if you're doing mass migrations uh your servers are replicated into aws what that means is that our agent uh reads every block and every disk that you chose to replicate and replicates it into aws and i'll talk a little bit later and uh time will also touch on this on our architecture and how it works so um and now i'm just doing a quick walk through uh once we finished copying that uh the initial data what we call the initial sync phase you can start testing uh how your servers run as ec2 instances so you you use the engine console to launch your servers is ec2 instances while this happens your source is uninterrupted the data replication process itself is uninterrupted and you know what a meaningful test is for you so we launched these as ec2 instances and you check that they do what they need to do so you're sure that when you do this again for cut over the cutover is gonna go smooth then again when you're ready for cut over you simply repeat that launch process for cutover purposes make sure that everything again works but you know that it's linked to because you've already tested it and you can do unlimited tests and non-disruptive and within minutes uh your server is running and then you can perform the actual cut over of directing traffic to your new servers um mgn is based on field proven uh technological pillars of calendar as i mentioned first of all is the continuous block level data replication that uh ensures that the differences between the source and the target are very small they can be sub second to be conservative let's call them seconds um it also knows how to continue so we're dealing with replication over the internet most typically that even though you can replicate without access to the public internet uh so it knows how to resume from where it left off without any user intervention it uses low cost uh ec2 and evs so while you're replicating you're not paying for that full server on aws you're only paying for a low cost replication server that's the ec2 and low cost ebs volumes for so for every disk you have the source there is a disk on aws but one replication server can uh deal with a replication of multiple source server so this is all this stage of the process is very low cost you only launch your actual first servers when you want to test or when you want to actually cut over and you can have a lot of control over how other replication works through replication settings which i won't dive into because i know every time we'll touch on this uh during his section of our talk today the second pillar is the orchestration i've already mentioned how we launch replication servers in order to facilitate the data replication so we do this automatically uh and we also terminate them as needed as i said every replication server can support multiple source servers so as you add more and as you finish cutting them over we will add and remove replication servers as needed and the same thing goes for uh when you launch an actual instance uh whether for testing or for uh actual cutover purposes uh we orchestrate all of that and you don't need to do any of that manually you don't need to go into the ec2 console or anything and the third pillar is the conversion this is really the magic source of the cloud android technology how we can take a server which is running on a physical server or any sort of virtualized hardware and make it work on as an ec2 instance um the next thing i'll touch on is um where we fit into the migration journey so we know migrations can take a long time it can take from weeks to years depending on what your needs are how many servers you're migrating uh with the technology and the architecture is and if you have compelling events such as data center exposure that's often the driving factor and often there are things that uh constrain the migration such as regulatory issues or other constraints so typically customers go through an assessment phase figuring out what do they have um and what is it that they actually want to move because customers don't always move everything to the cloud the next stage is readiness and planning making sure the servers that are going to be moved to the cloud are ready uh if they need to be updated anyway or etc and also planning the migration especially if you're migrating a large number of servers you're probably not moving them all at once and there are other planning actions then there's the execution which is what we do and then there's operation and optimization after the servers are already on aws and again the big advantage of using a lift and shifter re-hosting is you get your servers working quickly without any hassle on the cloud and then from there you go on and you optimize so mgn helps you execute the migration quickly and successfully uh mgm came out in ga in april on april 7th uh that's 2021 and we're now available in three u.s regions three europe regions and three uh regions in the asia pacific uh we know from experience with platinumer that discovers most of the needs for migrations but if you don't see the region you want to migrate into appear on this list uh we're not we are hard at work at ending all the other regions uh we're committed to being available in all aws regions and uh i'm going to now go and do a very quick demo just uh touch on a couple of points that have to do with uh uh making the best use of the user interface and after i'm done with that uh i'm going to hand it over to pretend to do a full hand um for walkthrough of a migration so here we're seeing uh the what we call the server list view in the uh mgm console we're seeing here a list of servers i'm using mock data here so uh if you're seeing the word preview or you're seeing any information that seems nonsensical don't worry about it uh the motivation of using mock data is for me to be able to show you some of the deep states um of the ui and really touch and sing some things that uh if i didn't point them out to you might take you a while to learn so generally speaking even though we're looking at a table and it's a list view this is basically uh designed to work as a dashboard when you're looking at your servers what you do is you look at this column if there's nothing in this column nothing requires your attention so you can just scan with rise in these notes which are color coded and there's only three of them with time you very quickly learn what they mean you don't even need to read the words that are in there but we'll stop let's talk about this one okay so this server um what would normally appear here by the way is the host name of the server so it will be very obvious to you which server so it says that it stalls stone means the data is no longer flowing and and and that this requires user intervention without user intervention this will not be fixed this is not those cases where we can there was a network hiccup and we can recover from it on our own often very common use case is when one person set up their application and someone else is responsible for the network shut down the ports that need to be opened for instance but you want to now find out once you run into an item that requires attention each row becomes a mini dashboard for that server so you know the host name you know that installed main data applications you see that it's not ready this is part of the migration life cycle that i'll touch on in a minute so you'll know what not really means right right basically means it's not ready to be launched tested or cut over yet when you go on reading you're saying okay we're in the initial sync phase and we've already might replicated 92 but we're stall now and what we have here is we always have a next step we always try to tell you what is the next thing that you need to do with each other in this case we need to resolve um the problem so we're going to dive in and try to find out more about what's going on with the server so here you can see we have a lot of tabs with a lot of information about this particular server but always at the top we tell you what your next sections are and show you any alerts if you need to in this case you need to uh resolve the problem of this uh that's causing the stole data replication and you can jump and see details it'll scroll right to the bottom provide you a lot of information which i'm not going to dive into right now what all of this means but what i want to tell you is we'll always tell you what the problem is it we'll also give you a link that will take you straight to the documentation page to how to solve specific problems because most issues that cause data or replication to be stolen are easily fixable on your own and you don't even need premium support you just need to drill into that server click on learn more find out how i can solve this and normally you'll be able to solve the problem yourself i'm going to go back to the list view now and i'm going to show maybe just one or two more things as i said i'm going to look now for a server that's in the state where um it's uh it's testing so this is a server where uh we can see we've already replicated 100 meaning we've completed that initial sync um so it's uh already passed the it derived at the ready for testing stage we've launched meaning we went to test we went to this menu and we've launched the test instance and now we're showing that the test is in progress so i've talked about the life cycle before this component is very important ties in also to the actions you can do here these all communicate with this it ties into the next actions that you can do here as you can see we're telling you here that you have a launch server the reason we're telling you this word well this is a note it's not a yellow or red notice that something we tell you about is because a launch instance costs you extra money we want to make sure you don't keep it launched if you don't need it right now and what it tells you here is that your next actions are to complete the testing and then to market is ready for cut over so what is to complete the testing you now have it running ec2 in this and you know what a test is meaningful for you it could be just to see that it launches an ec2 instance it could be that you want to ssh into it and see that you can connect to it and it could be that you want to see it operating in the context of a full uh application you know what a test is for you once you know that the test is done and good to go from this menu you choose mark is ready for cutover that will change the life and it will stay for ready for cut over and then when you look at the list view and i'll show you that again in a minute you can filter for servers that are in each state when you want to do certain but let's say the test didn't go through and you know you need to find out maybe change something on your source server or just explore this more you can revert it back to the ready for testing stage and the same thing goes with the cutover staging i'll also show you that while you're testing uh basically what we did is when you started the test again when you went here and you did launch test instance we did we launched an es2 instance that matches your source server and you want to know what's happening with that launch so you need to go to the ec2 console we show you right here we show you that it's launched we also show you if it arrived achieved the first boot meaning if it ever achieved two of two so on some servers especially windows servers that takes a little longer than just simply launching them so we track that and we show that to you here so you don't have to keep going to the ec2 console but if you want to we provide you a link to go right in there um when you launch this is a process that takes time and go through multiple steps i've talked about the conversion before so we use asynchronous jobs to do that so you can link right to the job from here that'll take you to this section which is called launch history i won't do a deep dive on that right now in this talk i want to leave plenty of time for the actual walkthrough and mostly i want to show you that you get you you can see a lot of information here uh i i've talked about the replication settings before so here uh you can see what all the settings are you can see there are quite a lot of them you can also see that you can edit them in this case we're adding them for a single server but you can also edit them for multiple servers at the same time i'll just maybe finish this demo by going to the source servers and showing that this uh this search bar here that appears in many of the new aws consoles is very powerful and it's very powerful here uh in particular first of all when you open it it already shows you what you can filter by but if you just want to see let's look for the word test okay so we see in this case we have a lot of host names with the name test and then but we can look for my servers who are ready for testing and now i'm seeing all the servers that are ready for testing um and the same way you can very easily filter your list um anything you want you can of course also uh sort this list by this and i'll make a note that when you sort it by migration lifecycle we're not sorting this alphabetically we're sorting this according to the same order that you saw um in that uh life cycle before going from not ready to ready for testing testing progress ready for cut over cut over in progress and then eventually cut over and eventually you can also cross the server to not be shown here anymore because if the cutover happened a couple months ago and you're working on you know your wave two three you're weighing number ten you probably don't want to keep seeing the server but even when you do that uh there's still a way to get and see your history i'll just quickly mention we have so this yellow alert which here it says lagging yellow basically means user intervention may be required lag means that the current state of the source server is uh more than those sub second or few seconds we said between that and the state are in aws so we recognize that and show you that and then we tell you that this is a problem that may resolve itself but also it may not anyone might want to explore that uh i think this is where i'll end uh my demo today and uh we'll move on to the walkthrough thank you yuron uh so let's get started with the demo here uh first things first uh so the first step in is ensuring mgn is initializing the region that you're planning to migrate so if you are not sure whether the service is initialized you can simply go to the service itself in the particular select the region that you want to basically migrate the servers to once you come here the home page you can click get started and that should actually tell you uh whether you have the service initialized in that region or not so in this particular uh region i have the service initialized so like for example if i select some another region here let's say singapore okay so as you can see i was redirected to setup application migration service so that's the that indicates that this application migration service is not initialized in singapore region so first things uh first is you would select the uh create replication settings template once you say create template this automatically initiates creating the required i am role for data replication and for the service to actually migrate your on premise virtual machines or servers to aws in form of ec2 instances now i'll jump back to ohio region for this demo now let's get started with im credentials because the first thing in order to get the uh as you don't mention the first step is installing the agent on the source machine uh for that we need to create an imuser which is which will provide uh with the required uh with the required permissions to uh to add the so add the source machine or source server to the aws console and from there complete the life cycle so first things first we'll go to the im console select add user now you can provide a username of your choice in this case i'm just providing mgn demo i'm gonna select programmatic access because everything will be done through command line here you can click next uh you can add to a specific group i'm gonna just attach a policy directly since uh i have the i have the service enabled so i do have the rules created so i'm gonna attach an existing policy so the policy that i'm attaching right now for this demo is the full access policy i'm just going to select that go to tags uh if you would like to add some tags for this particular user you can go ahead i'm going to skip that step and just going to quickly review uh what i have and say create okay uh i have the credentials created make sure that you save this file for future use because you cannot retrieve the credentials again you'll have to reset them now going back next step is going to be next step is going to add the servers now for the purpose of this demo uh we'll perform migration from us east 1 region to us east 2 for both windows and linux machines in parallel we'll also migrate a windows machine from vcenter to aws so to get started with let's go to the windows server that i have running in us east one okay open powershell run as administrator first thing is we are going to download the replication agent uh on the uh server so what i'm doing here oops sorry okay so in this scenario i'm using a invoke web request powershell command to download the uh installer as you can see in the url mentioned i'm using us east one s3 uh so the idea is whichever region is closest to your source server use that to download the installer so it's faster let's say if you're in usc 2 modify the url accordingly so let's start download here real quick okay while this is downloading i'm going to quickly get the credentials that we need for this okay we have the installer downloaded now i will just navigate to the folder where it's downloaded and execute the installer okay so the region name is going to be your destination region in this scenario i'm migrating to usc's 2. okay so this is the place where you will enter the access key id for the user that you have created and also specify the secret key okay so first thing after you enter the secret key is it's going to verify if it if the server has a minimum 2gb free space that is required after that it's going to start identifying the volumes for replication in this particular step you you have the ability to enter the volumes uh that you're going to migrate and for this particular machine i only have one volume so or so i'm not going to really enter a volume name but let's say if you want to migrate uh or replicate all disk all you have to do is simply hit enter and it's going to take the select all disk so as you can see it has identified the disks to replicate which is c the size is 30 gb and the agent is now downloaded and currently being installed while that is uh going on what i'm going to do is i have a linux machine in this scenario i'm just going to reconnect to it okay sorry about that so i have the linux machine here uh just gonna okay we will start the installation procedure uh it's similar to what we do on uh windows where first you will download the agent uh and then perform the installation okay so the agent is downloaded scene as you can see it has asked again uh what this q would i like to replicate i'm going to select all and the downloading of the agent has begun and now installing let's jump back to the windows and see what's the status in this scenario uh after installing the agent uh the source server was added to the mgn console and this is the server id in case you need to track it note it you necessarily don't have to keep a record this information is visible in the mgn console after this you can just hit enter and uh this window will be you'll you'll go back to the uh command prompt now let's jump to the console so as you can see in this scenario our windows machine is added and click on it real quick currently uh the replication initiation steps have begin so first thing is it's going to create the security groups required launch the replication server boot the replication server authenticate it download the required replication software uh create staging this now what this means is based on your source machine uh source server uh it's going to id it's going to create ebs volumes in your staging area identical so for the purpose of block level replication attach them to the replication server pair the replication server and the replication agent what this means is uh your source server is gonna get paired with replication server and they are going to basic source server is then going to start transferring data uh once the connection is finished the pairing is finished the data transfer begins now while this process is going on we will jump to the vcenter environment and perform a quick installation okay so i'm doing on a 2019 windows box here okay let's okay so in this scenario i have uh downloaded the agent already on this box ahead of time now we'll just execute the command to install the agent so another way to perform the installation is let's say you don't want to type the uh region or the access key or the secret key ever like it's step by step you just want a no prompt i execute one command and done you can certainly do that uh so as you can see i'm running the installer region access key secret key and then i've selected no prompt what that means is it's going to take all the disks automatically attached to this machine and start replicating okay there you go okay in meantime let's go back while this installation is finished oh okay sorry there is a typo in the command a child correcting as we speak i'll just wait uh okay there you go now it has it's working okay let's go back and check the status let me reload the page the session has expired i'm pretty sure many of you who are using single sign-on this is a very common experience that you may have after a certain time your session gets expired and then you have to refresh the screen and go back to start where you had stopped okay so one moment going back to the application migration service console okay as you can see uh i have the windows and the linux servers in here so the replication has started it provides you with a uh how much time is left again this is uh an approximate time everything depends on the network uh between the source and the destination region so make sure to check same with linux completed it has 11 minutes of course because the disk size is very small that it's eight gigs so it's pretty fast okay still creating the disk so i'll wait for that now in meantime while uh we are waiting for the replication uh to finish uh i'm just wanted to go over the architecture for the service real quick so this is a a network uh or a architecture diagram that you will find in our public documentation uh for mgn now i've just basically took this and doing it in an edit mode so we can have a con dive deep into it so let's start with this uh what this diagram means so on the left side we have our corporate data center so that could be vmware or you're doing migration from one aws region to another or any other cloud provider or so this is on your left is your data center or source on your right side we have the aws cloud uh specifically more your aws region where you're planning to migrate so as you saw during installation i had selected usb2 so this is my usb2 and this on the left is vmware slash usc one since i'm doing my replication oh sorry migration for three servers so first step is uh installing the agent which we did now to install the agent uh there are three main requirements or to successfully start replication and installing agent there are three requirements from a network standpoint the first is the source machine or source server should be able to talk to the mgn regional endpoint uh on port 443 tcp port 443 again you saw we were downloading the agent uh installer files uh through s3 so that means we also need access to s3 on tcp port 443 from the source server now the actual data transfer happens on tcp port 1500 so uh outbound uh tcp port 1500 needs to be open to the staging area now what is this staging here so before setting up mgn uh you should create a subnet which will be used by mgn as a staging area for data replicated from your source server to aws so if you remember at the start of the demo or the walkthrough we saw that there if you have not initialized the service you're asked to create uh or set up your replication template this is where you we suggest that you create a subnet for your staging area now you must specify uh this subnet in the replication settings template uh which you can create or modify for your server as i said while you can use the existing subnet in your aws account the best practice is to create new dedicated subnet for this purpose so there is no confusion around okay which subnet am i using of the existing if you need to modify something on that subnet it may end up impacting your migration journey so to avoid all that just create a new subnet up and associate that to the replication replication template uh moving forward with the network requirements as i said so the replication server was launched by mgn in your staging area subnet need to be able to send data over tcp port 443 to its regional api endpoint and s3 so we're talking about this traffic going from the staging area subnet to the mgn and s3 in addition uh the staging area also requires uh access to easy to end point of the region for the purpose of easy to resource management so for example we're talking about uh creating snapshots launching instances terminating instances cloning uh so cloning various up the aws resources the security groups up so for that purpose in addition the source server on which the replication agent is installed need to send data over tcp port 1500 uh to the replication servers and staging area as we discussed further the source servers will also send data to mgn regional uh api endpoints for example you can the one of the endpoint could be mgn dot u s e two amazon aws.com just to give you an example okay now we're talking about operational subnets which so this particular uh portion of the architecture which is which you can see is called migrated resources so when you do a test and cut over of the instances of the servers these are launched in subnet that you specify in your ec2 launch template associated with each source server so those uh who are familiar with cloud and your uh you may have seen this in form of blueprints uh those who are not uh or those who are new to mjn or what that means is uh in easy to launch templates you get an option to pick what settings you want for your target uh machine a target server target easy to instance so once you migrate the mesh the server in aws what instance type would you like uh what vpc what subnet uh what security groups uh get associated with that all these uh variety of uh configurations you are able to select now off after that uh let's let's discuss more along the lines of im roles now mgn is managed with two sets of impermissions one is the i'm credentials of the logged in user and the second is the use of service linked roles the this this allows uh grand control over permission so we are going to discuss a few uh service linked roles first we'll go over is the aws application migration service role policy now this policy allows the app mgn to manage aws resources on your behalf similar to what you may have seen from a cloud india perspective the cloud endeavor user used to do now the second is the aws application migration conversion server policy this policy allows the application migration service conversion server which our ec2 instance is launched by mgn to communicate with the mgn service so this is the this is this is where the conversion server is basically uh making your source server compatible with aws ec2 instance so whatever needs to be done at that point so in like for example installing aws drivers is one of the step that goes so licensing license activation that's part of the conversion process next is the aws application migration replication server policy this policy allows the mgn replication servers which are used uh which are ec2 instances launched by mgn to communicate with mgn service and to create eba snapshots in your aws account so the ebs snapshots are basically the data that is transferred from your source server to aws uh staging area subnet so in that so snapshots are taken so there is no data loss the next is the aws application migration agent policy this policy allows installing and using the replication agent which is used with aws mgn to migrate external servers to aws attach this policy to your im users or roles whose credentials you provide when installing the aws replication agent so this uh if you want to provide more explicit permissions to the user the imuse you have created you can select this particular policy as uh associate now let's see if you want to integrate it with migration hub there is a linked role for that as well which is called aws application migration mgh access so this policy allows aws mgn to send metadata about the progress of servers being migrated using mgn to aws migration hub the next one is the aws application migration full access which is the one i have uh associated the user to which i have attached to the im user so this policy provides permission to all public apis of aws mgn as well as permissions to read kms key and ec2 information so if if you decide to you encrypt everything in aws this is this policy can be used the next one is uh aws application migration easy to access this policy allows amazon ec2 operations required to use mgh to launch the migrated servers as ec2 instance again uh all these policies as you can see so far uh provides with granular access to the service so if you if you have a team uh who who is divided into okay certain members are going to install agent certain members are responsible for the actual testing and the remaining folks are responsible for the cut over you can provide granular permissions based on that now let's go back to the mgm console and see what's the status of our servers up let me go okay so as you can see the linux machine is fully migrated and as per the life cycle it is ready for testing the replication progress is 100 now let's go ahead and click on launch test instances so this provides a a small disclaimer you can say uh where it provides information what exactly is going on so we are about to launch easy to instance for one server uh as we selected uh these instances will be launched according to the launch settings that you have configured launch instances accuracy recharges as you know uh once an ec2 instance is in a running state it's gonna charge uh what the building starts on that so i'm gonna select launch and then while the instance is launched we will go back to the launch settings real quick so this is the mgn job id i'm just going to open this real quick okay so you can check the job log and track what exactly currently is going on uh so we'll let this process uh continue and we'll jump back uh after a few minutes so going to the launch settings real quick so there are two two things uh on this particular page one is the general launch settings so in this let me just edit real quick you can select the uh what you want as the instance uh state uh do you want to start the instance upon launch to stop uh you want to copy the private ip uh of the source and not uh you want to transfer in tags uh what kind of licensing you want so this is the basic lawn setting now a more granular launch setting would be or can be done via the ec2 launch template where let's go ahead and modify it real quick so every time you change the ec2 launch template a new version of the template is created and in order for mgm to use the new newly created version you must mark it as default so we'll click on modify this should automatically take you to the launch template console which is basically the ec2 management console okay so i'm just gonna give it mgn demo the name of the description okay in here i to get a chance uh to start with okay the main few things important you know you get a chance to select what instance type you want launch maybe uh c4 to x large is uh you feel is too bulky or uh you don't really need that kind of uh vcp use and memory so you can definitely go back and select let's say d3 for example let's select e3 large next is the key pair login we can select that vpc uh you can go ahead and select the vpc if you are using ec2 classic that's again an option uh volumes you can edit what kind of volumes you want the default is io1 you can select from a list available here gp2 gp3 which give us suits the needs here are the resource tag now these are the default tags that get created specifically the mgn tag you can see it's it has the server id that we saw during our post installation uh the name tag is what i gave to the instance and the volume so that gets carried over uh next is the network interface uh configuration you can go ahead and set up uh as per your needs a little more advanced stuff is if you want a i am policy or i am role attached for example if you are planning to use systems manager uh for this instance you can associate the policy i am role to that the same shutdown behavior what you wanted to be stop start hibernate behavior termination policies so as you can see there's a long list i'm not going to jump into each and every one but you get an option to select uh a variety of configurations for uh for launching and uh for launching the target instance i'm going to cancel this for now uh but did it and then we'll go back real quick and see what the progress okay so as you can see uh the task is completed successfully and a test instance is launched we'll walk through the steps real quick uh once we initiate the job it takes a little snapshot from the staging area subnet for the eps volumes where the data was replicated once that is completed the conversion process starts once the conversion process is complete uh we initiate uh the launch of target easy to instance now i'm gonna take this instance id and check it real quick okay again the session timeout so i'm just gonna reload so okay i am going to go to usb2 since that's the destination region for this migration okay uh as you can see here this is my target machine now we can confirm real quick okay so the instance id is ending in eight zero f five eight zero f five and you can see the status checks is two of two uh meaning it was a successful uh target launch now going back here uh now that you know this time you can go ahead and perform the required tests that you need to perform on your destination uh or your test launch from here as you can see test in progress so this is time to go ahead and test now uh i don't really have uh anything to test i just wanted to make sure that it passes status checks now next thing is i will be marking the server uh for cut over i made sure everything is good next step mark whatever it gives you again a little disclaimer here that test instances continue to occur easy to uh charges until terminated so it is recommended if you have completed the required test go ahead and let engine take care of terminating those instances and volumes associated with that so next time you launch you're not doubling your charges so you can select continue here and i've selected this option yes terminate launch instances okay so as you can see it's ready for cut over and you go back to this window okay as you can see the mgn is already shutting down the machine and it will be terminated shortly okay going back to my linux machine okay it is ready for cut over and i'm gonna start with the launch cutover machine so it's the same uh disclaimer as you saw for the test uh but this we're doing a launch so we'll do a launch now let's say you don't know the job id or you don't have that link you you do it from this screen so it's simple you can go to the launch history on the left side and it shows you uh which based on the date time stamp so as you can see started a few seconds ago so i can just go to this particular job it provides you with the start time and it has started again taking the latest snapshots so in meantime we'll go back to the okay windows machine is about eighty five percent done eighty-five percent done so nine minutes left uh check on the vmware source machine okay it is almost done as you can see it's creating first launchable snapshot what this means is data is replicated and a final snapshot is being created so that you can use the snapshot to launch uh a target machine for test or cut over as you can see so you can ignore the these details particularly because i was previously testing this in my lab with various machines so this machine was one of them uh where i was doing few test migration tests so and i later i uh stopped the migration reinstalled the agent just now for the demo so it basically keeps the history of the last test okay we will go back to the did we select the i'll just open launch history in another tab so it's easy okay as you can see the snapshot is finished the conversion has started go back i just want to make sure that we have at least one windows machine uh we can test out so i'm gonna monitor it so okay this is about 93 percent done uh both are in the phase of initial replication for initial sync mean time we will see the conversion started okay okay in meantime let's go over a few details um what we see for each source server so if you click on the server info you have the last update when was this particular server added what is the server id so this is you can say as the aws id of the arn for the server now once the agent is installed it collects various details so in this scenario you can see uh the cpu information how much memory uh tag for disks we will check the disk settings tab uh based on the cpu and memory of the source server mgn makes a recommendation what instance type should be used to launch the target is easy to instance uh you can of course change this if you if you think c4 2x large is uh you don't really need that kind of a instance type you can as i said you can go to launch settings then modify that so the host name is uh taken uh directly by the or is grabbed by the replication agent so mgn mln since this is a machine in us east one region we are doing u.s east one to usc's two so it provides way with the sources id uh its network interface uh information and what type of uh machine it is so it is a linux machine with all these details here next is the the tag so i have a name tag for that particular server so you can see it's engine demo going to the disk settings uh it provides you with the list of disks that are being replicated you can always come here and change the staging this remember this is not the target ec2 instance uh sorry target ebs volume this type this is the staging so what kind of this type you want to use for your replication environment um you can come here and select and this is the replication template that we were talking about so as you saw uh in the singapore region at the start of the demo i showed that uh it had asked to set up the or initialize the mgn servers this is the replication template uh which was displayed uh you can come here and modify the replication template later if you want to for each server this option is available you can come and go ahead and do it okay let's go back here and see okay so we do have uh a successful cut over for this particular machine the instance id is ending in nine eight ac zero so we will go to the so this is i'm gonna refresh the tab real quick so as you can see the previous uh launch instance is terminated and this is the new instance that is launch part of the cut over currently still going through the initialization state so it's we are waiting for it to pass the status checks but this is where you can come and check mean time we will see if we have a okay so as you can see we have uh the next step has finalized the cut over okay good so the windows machine is 100 in sync and creating snapshot as we speak let's just quickly check okay so there you go it is showing two of two status checks uh indicating everything is healthy from a server standpoint now this is going to be your final server so you have done a cut over now what we can do is go back to the machine itself and you can see the next step suggested is finalize the cut over so i'm going to select here and say finalize cut over so this is again a disclaimer where it says the action cannot be reversed this will cause all the replicated data to be discarded and all aws resources used for data replication to be terminated what this means is uh you have done the cut over you have this server in aws in form of an ec2 instance you don't really need to uh you don't really need the data transfer anymore everything is good and you are finalizing the cut over so once you click that the replicated data or is gonna get deleted so as you can see the data replication status change to disconnected and the cutover shows complete the next step is to mark it as archived so there is so you don't have a lot of servers here let's say this is you have like 100 servers here and you really don't want to keep sorting okay which servers i've already completed which are remaining so you can simply come here and mark it as archive sorry go here and mark as archived and it disappears so that's pretty much it and then if we go back the instance is still here so it did not so as you saw in the test launch uh you know after we said okay i'm ready for cutover it terminated but since we performed a cut over now it's not going to delete any of the launch instances because this is the final production ec2 instance that you have migrated okay okay so we have the vmware machine ready for testing uh based on the life cycles okay right here it has changed so i'm gonna go ahead and quickly do a launch test instance for this as euron mentioned uh earlier sometime windows machine may take some time because it has to go through additional uh additional configuration okay i will go to the launch history and check the m chain job okay as you can see just started taking the snapshots if we go back to the so when while this is uh while the rep the test launch is in progress everything that happens uh from a replication perspective or a test launcher cut over you can see the cloudtrail events where it provides you with additional details on the api calls you can use that for auditing purpose as well so right now it will sort the events just for this particular server id so which is sorry okay the server id is again ending in be5ee as you can see so you can select the resource name type it and it provides you with the details going back i'm going to see if the ec2 instance is ready yet it's creating a first launchable snapshot okay we will see okay so right now okay as you see uh it had previously launched uh as i said i had done some testing so it tried to clean up and the cleanup failed because i had manually deleted part of some test that was performing so it has moved on and currently doing the conversion uh for the easy twins for the vmware vm so while uh we are waiting for the chop the conversion to complete as you can see uh we performed three uh two sorry two test launch and one quarter so every time a new conversion server is launched by the mgn service so this so one is to one ratio as you can see there is a replication server running so going back to the architecture diagram we are talking about these staging areas up right here so we have the replication servers and the conversion servers being launched so you can come here and check that go back and check the job it is still um in the conversion phase so it will take maybe another uh five to seven minutes i would say at this point it also shows you how much time is uh how much replication time is elapsed uh so so far because as you don't mention uh the 90 days or the 2 1 the time that you can use the service free of charge is uh 90 days so you have uh every minute is calculated how much replication time so once it hits that 90 day convert it into minutes after that you will be charged for the replication okay we're still waiting for the conversion also let's take a look at uh in meantime if you go to volumes so you can see uh the volumes are tagged sorry not didn't mean to okay with so you have the mg demo up so all the volumes are created with the same tag as the uh ec2 instances and so it's easy for you to come to the volume section and say okay which are my uh which volumes are created by mgn uh and what's the purpose so it's easy to come here and track those same goes with snapshots as well right here it shows you uh sorry not this this uh it shows you that uh application migration services created the snapshot so it's easy for you to track that just going to quickly expand it real quick okay and as i have enabled encryption so you see all the snapshots are encrypted going back okay snapshot is still being created much too easy to instances dashboard in the meantime filter it by mgn so okay as you can see the source uh the conversion server is stopped so most likely the conversion is finished so we'll wait for the oh sorry another replication the conversion so let's go back to the okay so as you can see the conversion has ended and a this was the conversion server ids on line eighty i'm sorry so right here you can even track it this way which server is which conversion instance was used for uh target launch so it is currently in a stop state so meeting conversion has finished so now you're waiting for the target instance to be launched and it is already terminated so right here you can see uh a new machine being created and if you check the name so it is jnev if i go back so that is my uh vmware vm so convergence and it will just wait for the replication job to okay it shows completed status okay now let's go to the source server real quick and you can see here it shows as launched we'll go back to the instance there let's refresh this it's initializing as we speak okay we'll see if we can connect to it real quick actually not yet since we don't have the ip that means it's still being initialized so let's wait for a couple of minutes so one thing we can do is quickly check the status what's the instance doing and getting the instance screenshot refresh it okay i think it's still being processed okay so we do have the windows login screen now that means the instance has came up successfully so let's just wait maybe another minute or so so you're going to follow the same process that i mean the process is same whether it's a linux machine or a windows machine you're going to once the testing is complete uh you're going to click ready for cutover cut over cut over complete and then uh disconnect and archive the server at this point uh you can end the demo as the machine as you ca