none
Replication of HyperV VM's RRS feed

  • Question

  •  

    I'm currently testing a Hyper-V environment. I have two 2008 R2 Hyper-V servers with local storage. I want to replicate VM's at either guest or host level from one Hyper-V server to the other on their local storage (I may use ISCSI targets instead). If this had been VMWare I would probably have used Veeam but it isn't available for Hyper-V. I don't want to be tied to a 3rd party storage product like starwind, steeleye etc. I'm not sure I fancy Doubletake either. I don't have enough money for a NetApp etc. for lun replication.  In addition, I like the idea of a replication solution that is storage agnostic.

    When I say replication, it doesn't have to be real time but near real-time would be nice (every 15 mins or so). Recovery should be instant or near instant; i.e. server 1 explodes I can start the VM's up on server 2 with negligible data loss. I tested AppAssure Replay today and thought it worked quite effectively although you have to install agent at guest level of VM. Nevertheless it seemed to tick the boxes but I would like to know what else there is. DPM 2010 replicates but doesn't give instant recover, is that correct?

    Has anyone else on the forum used Replay 4 in a live environment? 

    Thanks in advance.

     



    Tuesday, March 29, 2011 10:59 PM

Answers

  • Most customers who want high availability and disaster recovery use a combination of solutions.  For example, you could move the VMs and VHDs from local storage to shared storage and create a failover cluster.  R2 allows you to leverage (CSV) Cluster Shared Volume “Live Migration” or classic quick migration were a VM’s state is saved to shared storage before failover.  This provides redundancy in the event that one node fails.  DPM can be used to backup the entire VM but also provides more advanced protection using Volume Shadow Copy.  DPM calls into VSS to get all the Microsoft workloads in a consistent state for backup.  Let’s say one of your VMs is running Exchange, you could protect the entire VM and the Exchange workload.  Therefore in the event of disaster you can restore the entire VM based on the last recovery point and the replay the logs for Exchange.  DPM essentially provides constant protection. 

    Dave Guenthner [MSFT]
    This posting is provided "AS IS" with no warranties, and confers no rights.

    http://blogs.technet.com/b/davguents_blog/

    Wednesday, March 30, 2011 12:44 AM
  • Hi,

     

    As Dave mentioned, you need to set up a Hyper-V Failover Cluster if you want to let the VM's up on server 2 with negligible data loss when server 1 is exploded.

     

    By the way, you mentioned “I may use ISCSI targets instead”, so if you have iSCSI storage in your environment, you can build a Hyper-V Failover Cluster without other storage product like starwind.

     

    For more information, you can refer to the following guides.

     

    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://www.microsoft.com/downloads/details.aspx?FamilyId=CD828712-8D1E-45D1-A290-7EDADF1E4E9C&displaylang=en

     

    Adding a Pass-through Disk to a Highly Available Virtual Machine

    http://blogs.technet.com/askcore/archive/2009/02/20/adding-a-pass-through-disk-to-a-highly-available-virtual-machine.aspx

     

    Understanding Cluster Shared Volumes in a Failover Cluster

    http://technet.microsoft.com/en-us/library/dd759255.aspx

     

    Cluster Shared Volumes Support for Hyper-V

    http://technet.microsoft.com/en-us/library/dd630633(WS.10).aspx

     

    Deploying Cluster Shared Volumes (CSV) in Windows Server 2008 R2 Failover Clustering

    http://blogs.msdn.com/clustering/archive/2009/02/19/9433146.aspx

     

    Hyper-V: Using Live Migration with Cluster Shared Volumes in Windows Server 2008 R2

    http://technet.microsoft.com/en-us/library/dd446679(WS.10).aspx

     

    http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf

     

     

    Best Regards,

    Vincent Hu

     

     

    Wednesday, March 30, 2011 6:42 AM
    Moderator
  • You should really take a closer look at the information that Vincent and Dave provided you.

    How many 9`ers do you have in your SLA? You would basically need the following:

    • Failover Cluster - to achieve high availability, this is the solution as far as Microsoft concern when it comes to the in-built features in Windows Server. You will be able to failover your VMs from node1 to node2 etc, in case of a hardware failover. In addition, you can also use guest clustering within your VMs - using iSCSI.
    • Some sort of storage replication. Most of the good vendors provides various solutions here (replication). And you can, as Dave mentioned, use DPM2010 to backup your VMs vhds every 15 minutes, so the downtime should be minimal. 

    Conclusion: A simple way to achieve HA may sound impossible. 

    To achive this myself, I have a failover cluster, 2 SANs, and 2 switches. The key word is to eliminate the single point of failure.  24/7 has its price.

     


    Kristian (Virtualization and some coffee: http://kristiannese.blogspot.com )
    Wednesday, March 30, 2011 10:12 AM

All replies

  • Most customers who want high availability and disaster recovery use a combination of solutions.  For example, you could move the VMs and VHDs from local storage to shared storage and create a failover cluster.  R2 allows you to leverage (CSV) Cluster Shared Volume “Live Migration” or classic quick migration were a VM’s state is saved to shared storage before failover.  This provides redundancy in the event that one node fails.  DPM can be used to backup the entire VM but also provides more advanced protection using Volume Shadow Copy.  DPM calls into VSS to get all the Microsoft workloads in a consistent state for backup.  Let’s say one of your VMs is running Exchange, you could protect the entire VM and the Exchange workload.  Therefore in the event of disaster you can restore the entire VM based on the last recovery point and the replay the logs for Exchange.  DPM essentially provides constant protection. 

    Dave Guenthner [MSFT]
    This posting is provided "AS IS" with no warranties, and confers no rights.

    http://blogs.technet.com/b/davguents_blog/

    Wednesday, March 30, 2011 12:44 AM
  • Hi,

     

    As Dave mentioned, you need to set up a Hyper-V Failover Cluster if you want to let the VM's up on server 2 with negligible data loss when server 1 is exploded.

     

    By the way, you mentioned “I may use ISCSI targets instead”, so if you have iSCSI storage in your environment, you can build a Hyper-V Failover Cluster without other storage product like starwind.

     

    For more information, you can refer to the following guides.

     

    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://www.microsoft.com/downloads/details.aspx?FamilyId=CD828712-8D1E-45D1-A290-7EDADF1E4E9C&displaylang=en

     

    Adding a Pass-through Disk to a Highly Available Virtual Machine

    http://blogs.technet.com/askcore/archive/2009/02/20/adding-a-pass-through-disk-to-a-highly-available-virtual-machine.aspx

     

    Understanding Cluster Shared Volumes in a Failover Cluster

    http://technet.microsoft.com/en-us/library/dd759255.aspx

     

    Cluster Shared Volumes Support for Hyper-V

    http://technet.microsoft.com/en-us/library/dd630633(WS.10).aspx

     

    Deploying Cluster Shared Volumes (CSV) in Windows Server 2008 R2 Failover Clustering

    http://blogs.msdn.com/clustering/archive/2009/02/19/9433146.aspx

     

    Hyper-V: Using Live Migration with Cluster Shared Volumes in Windows Server 2008 R2

    http://technet.microsoft.com/en-us/library/dd446679(WS.10).aspx

     

    http://h20341.www2.hp.com/ERC/downloads/4AA2-2983ENW.pdf

     

     

    Best Regards,

    Vincent Hu

     

     

    Wednesday, March 30, 2011 6:42 AM
    Moderator
  • I know I can have a cluster but I want to protect against the storage failing as well. Going back to the scenario where I don't have ISCSI targets and I'm simply using two Hyper-V servers with local disk array storage. One server runs my live VM's and I want to replicate to the other (with instant recovery). With local storage if one server fails I lose both hyper-v and the data so a cluster with shared storage won't work. Any other suggestions or am I missing something here.

     

    Thanks Again

     

    Regards

    Richard Abrams

    Wednesday, March 30, 2011 9:13 AM
  • You should really take a closer look at the information that Vincent and Dave provided you.

    How many 9`ers do you have in your SLA? You would basically need the following:

    • Failover Cluster - to achieve high availability, this is the solution as far as Microsoft concern when it comes to the in-built features in Windows Server. You will be able to failover your VMs from node1 to node2 etc, in case of a hardware failover. In addition, you can also use guest clustering within your VMs - using iSCSI.
    • Some sort of storage replication. Most of the good vendors provides various solutions here (replication). And you can, as Dave mentioned, use DPM2010 to backup your VMs vhds every 15 minutes, so the downtime should be minimal. 

    Conclusion: A simple way to achieve HA may sound impossible. 

    To achive this myself, I have a failover cluster, 2 SANs, and 2 switches. The key word is to eliminate the single point of failure.  24/7 has its price.

     


    Kristian (Virtualization and some coffee: http://kristiannese.blogspot.com )
    Wednesday, March 30, 2011 10:12 AM
  • Hello,

    In addition to Live Migration, Microsoft DPM can be used to backup the virtual machines, replicate them to another site's DPM server, then restore them to any Hyper-V host at the other site.  This is a great way to kill two birds with one stone.  You certainly can use the other technologies you discussed, such as redundant SANs, replicated storage, or third party products as well, though the price goes up substantially.

    Here is some more information on DPM:

    http://blog.concurrency.com/infrastructure/back-me-up-im-going-in-hyper-v-and-backup/

    Nathan Lasnoski


    http://blog.concurrency.com/author/nlasnoski/
    Thursday, March 31, 2011 4:06 AM
  • Hello,

    Did this address your question?  Thanks,

    Nathan Lasnoski


    http://blog.concurrency.com/author/nlasnoski/
    Friday, April 1, 2011 4:02 PM
  • Its possible replicate virtual machine w/sql to server in another site whitout SAN or ISCOSSI envirolment w software or anther tool.

     

    Tuesday, August 23, 2011 4:14 PM
  • There are 3rd party Applications for replication.

     

    1. Double take - http://www.visionsolutions.com/Products/DT-Avail.aspx

    2. Plate Spin - http://www.novell.com/products/protect/

    These both eliminate the need for a SAN and allow you to replicate your VM's.


    Gil Gross | Project Manager & Technical Consultant | G-Net Network Solutions | www.g-net.co.il
    • Proposed as answer by Gil Gross Tuesday, September 13, 2011 8:17 AM
    Tuesday, September 13, 2011 8:17 AM
  •  

    ...If this had been VMWare I would probably have used Veeam but it isn't available for Hyper-V.

     



    Quick update, Veeam Backup & Replication is now available for Hyper-V :)
    Monday, September 24, 2012 11:11 PM