Hyper-V on shared storage with failover cluster setup correctly?
- Ok. I am trying to setup two Dell PE2950 servers to be in a hyper-v failover cluster with all of the hyper-v stuff on a SAN. My goal is to use both servers to host VMs but if one server fails they will "live migrate" over to the other server. So I have 2008 R2 Datacenter on each sever. I made one of the servers, lets say server01, a PDC and created a new AD and then I had server02 join that domain. I then installed hyper-v and failover on each server. I then did a failover verification, and it passed, and then made a failover cluster. I then enabled Cluster Shared Volumes. Each server has 2x 1GB broadcom and 4x 1GB intel cards. The SAN has the same ethernet setup, iSCSI SAN. I made a VLAN and etherchannel for the 4x 1GB intels so the iSCSI traffic has a 4GB connection and it is on a VLAN all be itself.So that is what I have done so far. Now I have a question about the LUN setup. I assume I need to create a LUN for each hyper-V OS I want to install? So if I want to install 20 VMs, then I need to create 20 LUNs? Also what if I wanted to separate page files and data from OS? Do I setup a LUN for OS, PF, and DATA? Now after I have LUNs setup, I want to make it so that both servers can see the LUN, right? So I setup my iSCSI initiator on my windows servers to point to the SAN. So lets say I setup a 200GB LUN called TESTOS on my SAN. Now on server01 and server02, I can see this offline and uninitilized disk. So on server01 I initilized the disk as a GPT disk. Is GPT ok or should I do MBR? After that I then format the drive as NTFS and give it a name of TESTOS. I do NOT assign it a drive letter. I am watching server02 as I do this and the changes I make on server01 are showing up on server02. So now I go into the failover cluster manager.I add this TESTOS disk to the storage, then I add the TESTOS into the Cluster Share Volumes???? Is that what I need to do? This is the part I am a bit confused about. Once it is in the Cluster Shared Volumes it says that the status is online and that the owner is server01. Does that mean that when I create a hyper-v VM, I need to create the VM for this LUN on server01? Is there anything else I need to do here?Now onto the Hyper-V part. I create a new VM, I give it a name of TESTOSVM. When it asks where I want to store the virtual machine, where should I point it too? Should this be the same place for all VMs? When it asks for the virtual hard disk, should I point it to the TESTOS drive?Thank you in advance for your help.
Answers
Hi,
If you don’t use CSV on Windows Server 2008 R2(or you are using Windows Server 2008), the disk reserved on another node is an expected behavior. In a Cluster environment, one disk can be accessed only by one node at the same time, if you move or migrate the VMs to another node, you will be able to access the disk on another node. So we introduce Cluster Shared Volume, after you enable CSV on Windows Server 2008 Cluster environment, you will be able to access the disk on every nodes.
Deploying Cluster Shared Volumes (CSV) in Windows Server 2008 R2 Failover Clustering
http://blogs.msdn.com/clustering/archive/2009/02/19/9433146.aspx
Vincent Hu
- Marked As Answer byVincent HuMSFT, ModeratorThursday, October 15, 2009 6:12 AM
All Replies
- Hi,
This article explains how to setup shared storage with hyper-v for a failover cluster. Unfortunatly the article is written in Dutch.
http://hyper-v.nu/blogs/hans/archive/2009/01/03/configuratie-van-cluster-shared-volumes-csv.aspx
If you need translation for the article then let me know.
Danny.- Proposed As Answer byDanny van Dam Saturday, October 03, 2009 8:12 AM
Hi,
I assume I need to create a LUN for each hyper-V OS I want to install? So if I want to install 20 VMs, then I need to create 20 LUNs?
>> You don’t have to do this in Windows Server 2008 R2, you can put many VMs in a LUN, however, if you care about the performance, you can put one VM on one LUN.
Also what if I wanted to separate page files and data from OS? Do I setup a LUN for OS, PF, and DATA?
>> This is not related to Hyper-V. Now you can think the VM as a normal computer, if you want to keep the page files and data files from OS, you can do it as you did in a normal computer.
In addition, I add some articles about Hyper-V Failover Cluster, you can refer to:
Hyper-V: Using Hyper-V and Failover Clustering
http://technet.microsoft.com/en-us/library/cc732181(WS.10).aspx
Step-by-Step Guide for Testing Hyper-V and Failover Clustering
http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf
Important Note: This response contains a reference to a third party World Wide Web site. Microsoft is providing this information as a convenience to you. Microsoft does not control these sites and has not tested any software or information found on these sites; therefore, Microsoft cannot make any representations regarding the quality, safety, or suitability of any software or information found there. There are inherent dangers in the use of any software found on the Internet, and Microsoft cautions you to make sure that you completely understand the risk before retrieving any software from the Internet.
Vincent Hu
- Great articles. Thank you. I think I have it figured out. Now another question about how to setup the network access. I have a two port nic and a four port nic. I am using iSCSI. I was wondering if I should setup the two port NIC for normal management, two of the four ports for VLAN access to the iSCSI SAN, and the other two ports for access to my normal network. I could etherchannel them all so each would be a 2GB connection. Does this setup sound good or should I go a different route? The VLAN that my iSCSI SAN is on is all by itself with no other devices on it.
Hi,
Network infrastructure and domain account requirements for a two-node failover cluster
You will need the following network infrastructure for a two-node failover cluster and an administrative account with the following domain permissions:
- Network settings and IP addresses: When you use identical network adapters for a network, also use identical communication settings on those adapters (for example, Speed, Duplex Mode, Flow Control, and Media Type). Also, compare the settings between the network adapter and the switch it connects to and make sure that no settings are in conflict.
If you have private networks that are not routed to the rest of your network infrastructure, ensure that each of these private networks uses a unique subnet. This is necessary even if you give each network adapter a unique IP address. For example, if you have a cluster node in a central office that uses one physical network, and another node in a branch office that uses a separate physical network, do not specify 10.0.0.0/24 for both networks, even if you give each adapter a unique IP address.
You route is ok, however, one port NIC for host management is enough. We have another post discussed the similar issue, you can refer to:
Network configuration when clustering Hyper V
http://social.technet.microsoft.com/Forums/en-US/winserverhyperv/thread/023680f3-fc5e-4b45-ae17-fcb12f148f1aVincent Hu
- Network settings and IP addresses: When you use identical network adapters for a network, also use identical communication settings on those adapters (for example, Speed, Duplex Mode, Flow Control, and Media Type). Also, compare the settings between the network adapter and the switch it connects to and make sure that no settings are in conflict.
- How does the pdf below look for network setup? I took the suggestions from the post you linked about having iSCSI, cluster, management, and VMs to have their own NIC or NICs. I also segrated iSCSI and Clustering to their own VLAN.Hyper-v network config PDFhttp://www.jdover.com/hypervnetworkconfig.pdfNow when I create a VM, it asks where to store the config files. I moved them to a LUN called V:. Is that ok?When I create a VM, it asks about what to do with the hard disk. I told it to create a fixed disk file to a 200GB LUN called M:. Is that ok?I ask these questions becuase in the Step to Step Guide for Testing Hyper-V Failover, in step 6 it tells me to create a VM."1. Open Hyper-V Manager. Click Start, point to Administrative Tools, and then click Hyper-V Manager.2. If you are not already connected to the server that owns the shared storage, connect to that server.3. From the Action pane, click New, and then click Virtual Machine.4. From the New Virtual Machine Wizard, click Next.5. On the Specify Name and Location page, specify a name for the virtual machine, such as FailoverTest. Click Store the virtual machine in a different location, and then type the full path or click Browse and navigate to the shared storage.6. On the Memory page, specify the amount of memory required for the operating system that will run on this virtual machine. For example, specify 1024 MB to run Windows Server 2008.7. On the Networking page, connect the network adapter to the virtual network that is associated with the physical network adapter.8. On the Connect Virtual Hard Disk page, click Create a virtual hard disk. If you want to change the name, type new a name for the virtual hard disk. Click Next.9. On the Installation Options page, click Install an operating system from a boot CD/DVD-ROM. Under Media, specify the location of the media, and then click Finish.ImportantDo not start the virtual machine at this point. The virtual machine must be turned off so that you can make it highly available."So for part 5 and part 8, is pointing these to a LUN what I need to do? Because later in the document, Step 7, it makes the VM highly available by adding it to the cluster. If you follow these steps and do NOT have the V: and M: LUNs added to the Cluster storage, then the Failover Hyper-V VM test will toss up a warning saying the storage is not in the cluster.
- Have you tried the guide from HP: http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf , it's elaborate, I think you will get all the answers there.
Have you tried the guide from HP: http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf , it's elaborate, I think you will get all the answers there.
Ill check it out.- Hi,
Please take your time!
Best regards,
Vincent Hu
Have you tried the guide from HP: http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf , it's elaborate, I think you will get all the answers there.
I followed the instructions all the way to Step 7. When I add the V: drive to my failover cluster storage, the V drive goes into "Reserved" mode on server02 and no longer is mappable. Which means in the next steps, I cannot create a VM and point it to the V drive. The V drive shows up on server01 but not server02. IT tells me it is reserved.Hi,
If you don’t use CSV on Windows Server 2008 R2(or you are using Windows Server 2008), the disk reserved on another node is an expected behavior. In a Cluster environment, one disk can be accessed only by one node at the same time, if you move or migrate the VMs to another node, you will be able to access the disk on another node. So we introduce Cluster Shared Volume, after you enable CSV on Windows Server 2008 Cluster environment, you will be able to access the disk on every nodes.
Deploying Cluster Shared Volumes (CSV) in Windows Server 2008 R2 Failover Clustering
http://blogs.msdn.com/clustering/archive/2009/02/19/9433146.aspx
Vincent Hu
- Marked As Answer byVincent HuMSFT, ModeratorThursday, October 15, 2009 6:12 AM
· Hi,
Have you tried the suggestion? I want to see if the information provided was helpful. Your feedback is very useful for the further research. Please feel free to let me know if you have addition questions.
Best regards,
Vincent Hu
Vincent Hu,Hi,
If you don’t use CSV on Windows Server 2008 R2(or you are using Windows Server 2008), the disk reserved on another node is an expected behavior. In a Cluster environment, one disk can be accessed only by one node at the same time, if you move or migrate the VMs to another node, you will be able to access the disk on another node. So we introduce Cluster Shared Volume, after you enable CSV on Windows Server 2008 Cluster environment, you will be able to access the disk on every nodes.
Deploying Cluster Shared Volumes (CSV) in Windows Server 2008 R2 Failover Clustering
http://blogs.msdn.com/clustering/archive/2009/02/19/9433146.aspx
Vincent Hu
So should I setup my VMs on CSV? Will that cause things to run slower?Hi,
So should I setup my VMs on CSV? Will that cause things to run slower?
>> It really depends on your requirement. Generally speaking, if you are running Windows Server 2008 R2, we recommend that you enable CSV on Hyper-V computer, especially you are working in a cluster environment.
By the way, there is no feedback that CSV will cause the computer slower, actually, it will makes your VMs work faster.
Best Regards,
Vincent Hu
- Just want to point out one thing, you wrote "My goal is to use both servers to host VMs but if one server fails they will "live migrate" over to the other server." If one of the servers fails (at least if Windows Failover Clustering thinks it has failed) the VM will be brought up on the other node (assuming correct clustering configuration) but it will NOT be a seamless live migration or even a quick migration, it will be a startup of the OS from the last state stored to disk. For live migration or quick migration to work, both sending and receiving nodes have to be operational and cooperating.
thanks
Martin
