Chapter 5: Configuring Your Private Cloud Infrastructure
Overview
In the previous chapter we connected all the components of System Center 2012 together. Now, we are ready to configure the infrastructure components within our private cloud.
One of the key tenets of a private cloud is the ability to abstract your diverse set of physical resources into a logical pool that you can delegate to specific users. Jeff wants to be able to give resources to Emily so that she and her team can do their work. The problem is that Emily doesn’t have knowledge of the compute resources, the different internal network architecture or the storage topology. In reality, Emily shouldn’t be concerned with that.
Emily would just like access to a pool of resources to deploy the applications her team needs to deploy. Jeff wants to provide Emily standard choices for memory, storage, which network to connect to and the number of virtual CPU’s so that she and her team can easily create a virtual machine and deploy applications without the need to know the underlying infrastructure details.
Jeff is going to prepare the private cloud infrastructure now through Virtual Machine Manager. He needs to import the existing Hyper-V host (or hosts, depending on your scenario) and create a network private cloud infrastructure called SEATTLE_EXTERNAL. He also wants to explore some of the other fabric possibilities such as the storage classifications.
Private Cloud Infrastructure Components
When working with the private cloud, it is very important to understand the concept of Infrastructure
Components. In the past, most of the datacenter administrators talked in terms of hardware. With the private cloud, we are going to abstract the hardware into pooled resources. Components such as networking, storage and compute (CPU, memory) will be abstracted from the hardware and delegated to Emily. It is Jeff’s
responsibility to manage the underlying physical resources and just ensure that Emily has the correct amount of resources available to suit her business needs. Emily does not need to know or understand the underlying hardware to be able to do that.
Host Groups
Host groups are used to group virtual machine hosts in meaningful ways, often based on physical site location and resource allocation. Jeff has two datacenters and is going to create a specific host group for this evaluation, and create a structure that matches the Brussels Datacenter. For each group, Jeff can define different parameters based on the infrastructure that will be placed beneath that host group.
Jeff starts with creating the Seattle host group.
1. In the Virtual Machine Manager console, go to VMs and Services > All Hosts
2. On the ribbon, select Create Host Group
3. Give it the name Seattle
4. Jeff also is going to have a quick look on how he can create a tree for a datacenter that Contoso has in Brussels. Below you see the examples.
5. Jeff realizes that he can do some great things with the host groups. However, because he is still
evaluating the private cloud, he is going to keep working with the Seattle host group. Jeff right-clicks on host group and select's Properties to adjust some settings specific for that host group.
6. On the General page, you see the name that you gave the host group when you created it. You can add a description here, change the location if necessary, and decide if you want to allow unencrypted file transfers. Type in the following data:
Description: Main host group for the Seattle Datacenter
7. On the Placement Rules page, you can define specific placement settings for this host group
8. On the Host Reserves page, Jeff decides to change a few settings specific for the Seattle hosts. Change the settings to the following parameters:
CPU: 5 %
Memory: 256 MB
Disk I/O: 0 IOPS
Disk space: 1 %
Network I/O: 0 %
9. On the Dynamic Optimization page, Jeff looks at the specific options that he has to balance the virtual machine load within a host cluster. Because Jeff doesn’t have a cluster in this evaluation it isn’t that important, but it is good to see that he has many options to work with. He also selects the Settings button to look at the power optimization settings.
10. After looking at the Power Optimization settings, select OK
11. On the Network page, Jeff doesn’t see any network resources yet. But since we still need to create the private cloud infrastructure, they will appear here later on.
12. On the Storage page, you can review the Storage allocated to this host group. Again, since we didn’t create any private cloud infrastructure yet, we don’t see any storage here.
13. On the Pro Configuration page, you can view the settings for the Pro Monitors. You can overrule the inheritance and change the settings specific for this host group.
14. On the Custom Properties page, you can manage Custom Properties for your hosts.
Networking
Jeff now has some host groups created. In his evaluation, he is only going to use the Seattle host group and now he wants to create a logical network for that host group together with some predefined IP addresses that can be used when deploying the servers for Emily’s application.
Create Logical Network
Jeff is going to create a network in Virtual Machine Manager. This network is called the SEATTLE_EXTERNAL network. When your server is connected to this network, you will get internet access.
It is perfectly possible that you don’t have a specific network. In our evaluation, we only have one network on the host and that network will be imported when we are going to import the hyper-v host. The next exercise is just to create a feeling about how to create a logical network and what the possibilities are 1. In the Virtual Machine manager console, go to Fabric > Networking
2. On the ribbon, select Create Logical Network
3. On the Name page, fill in a name and description for the network. Please note that a well-chosen name and description are crucial here. Don’t forget that it will be Emily that will chose some of these private cloud infrastructure or resources so she will need to know what network it is and what the function of that network is. Select Next
• Name: Exterior – Seattle Network
• Description: Network that gives access to the Internet
5. Select the specific host group (Seattle) and fill in the IP Subnet in. For this example, I’m using the 10.0.0.1/24 subnet and no VLAN. If you have a VLAN, you need to fill it in here. Also, you can change the Network site name from the suggestion. Select OK
6. On the Summary page, confirm the settings and select Finish
Create IP Pool
Now that Jeff has created a logical network he is going to create an IP Pool for that logical network.
1. In the ribbon, select Create IP Pool
2. On the Name page, select the Seattle Network and type in the information for the IP Pool. Then select Next
Name: Internet Seattle Pool
Description: All IP-addresses that can be used to connect to the internet from the Seattle Host Group
Logical Network: Exterior – Seattle Network
3. On the Network Site page, select Use an existing network site and make sure that the correct Seattle Network is selected. Then select Next
4. On the VIP and Reserved IPs page, fill in the starting IP address and ending IP address. This will give you your total number of IP addresses that can be used in that pool. You can also specify specific IP addresses reserved for load balancers or for other reasons. Jeff fills in the following data as example and then select Next
Starting IP address: 10.0.0.10
Ending IP address: 10.0.0.49
IP addresses reserved for load balancer VIP’s: 10.0.0.25
5. On the Gateway page, you can specify a gateway for that IP pool. Jeff fills in the IP address for the gateway (10.0.0.1) in that network. Select Next
6. On the DNS page, you can fill in the IP of the DNS server(s) for that network. You can also specify a DNS suffix. Jeff fills in the DNS server 10.0.0.2. Select Next
7. On the WINS page, you can specify WINS servers. Jeff doesn’t have those anymore so he selects Next
8. On the Summary page, review your selections and settings and select Finish
Create VIP Template
Jeff decides to create a VIP template also as Emily will probably need a load balancer in the near future.
1. On the ribbon, click Create VIP Template
2. On the Name page, fill in the name for the template, the description and the Virtual IP port and then select Next
Template name: Seattle VIP Template
Description: The template is used for the webservers in the Seattle Host Group
Virtual IP Port: 80
3. On the Type page, Jeff selects to create a Specific type because he knows that he is going to use the Microsoft Network Load Balancing feature in the future. Select Next after selecting Microsoft as manufacturer and the correct Model: Network Load Balancing (NLB)
4. On the Protocol page, Jeff selects the TCP protocol and then he selects Next
5. On the Persistence page, Jeff leaves the Enable persistence open and selects Next
6. On the Summary page, review your selections and select Finish
Storage
Now that Jeff has configured the network private cloud infrastructure, he wants to explore the storage section as well. Unfortunately, because this is an evaluation he doesn’t have storage that meets the SMI-S specification (http://www.snia.org/tech_activities/standards/curr_standards/smi). Storage that is SMI-S capable is necessary to allow Virtual Machine Manager to automatically discover storage. So for this evaluation, he is just going to explore the classifications.
Jeff knows he that he will have to configure SMI-S when he goes into production. He reads about it on the TechNet documentation (http://technet.microsoft.com/en-us/library/gg610600.aspx) and sees the possibilities for more automation in the future.
1. Go to fabrics > Storage
2. On the ribbon, select Create Classification
3. On the Create a storage classification page, enter a name and description for the storage and select Add
4. In this example, Jeff has created classifications for Bronze, Silver and Gold storage as he will do when going into production.
Compute
Jeff has made all the preparations to start importing his Hyper-V server. This is the last fabric he is going to configure. Because the Hyper-V server already exists, all he needs to do is add the server to the Seattle host group. Virtual Machine Manager will automatically deploy the agent to the host and add that host as a managed entity in Virtual Machine Manager. Jeff doesn’t need to create the same virtual network that already is configured on the host or add the storage (which is local storage in this evaluation) as Virtual Machine manager will
automatically import the network configuration.
1. In Virtual Machine Manager, go to Fabric > Servers
3. On the Resource Location page, select Windows Server computers in a trusted Active Directory domain and select Next
4. On the Credentials page, select Use an existing Run As account and select Browse…
5. On the Select a Run As Account, select Create Run As Account
6. On the Create Run As Account, type in the Name and Description for the account. Because Jeff is evaluating, he choose his own account for now. In a real production environment, he will never choose his own user but a specific run as user. Type in the password for Jeff (twice) and select OK
7. On the Select a Run As Account page again, you can now select the Domain RAA account. Select OK
8. Back on the Credentials page, select Next
9. On the Discovery Scope page, select Specify Windows Server computers by names and type in the name of your host in the text box. (Hyperv02). Select Next
10. On the Target resources page, select the discovered server and select Next
11. On the Host settings page, select the Seattle host group, type in the path where you want to store your virtual machines (in our evaluation environment, this is a local path) and select Next
12. On the Summary page, review your selections and select Finish to add the host to your environment
13. When the job is finished, you will be able to see your host added to the Virtual Machine Manager console
14. You will also see that under Fabric > Networking > Logical Networks the configured network on the host is now visible
Summary
Jeff has created private cloud infrastructure. He also investigated the possibilities he will have in a true production environment. Now that Jeff is ready with all the preparations, he is ready to move on to the next chapter, and build a cloud for Emily and her team. In that chapter, Jeff will see the true value of creating a pool of resources.