Hyper-V Replica

With Windows Server 2012, Microsoft has introduced a great new feature in the ever growing Hyper-V environment.  Hyper-V replica is a game changer in terms of providing a built in replication technology that is included even in the standard edition of Windows 2012.

I have had a chance to do some testing with this new feature and have been quite impressed with the operation.  In my test environment I used to machines with local storage.  I was able to create a VM on the primary Hyper-V machine and establish a replication link to my replica server.  After configuring the VM for replication, the VM was configured as a replica on the replica server and a file copy began.  You can monitor statistics of the replication through the Hyper-V manager interface.  Replication is setup to be asynchronous.  At this time you are limited to a replication of changes every 5 minutes and an application consistent backup at often as every hour through VSS.  This option will ensure the application is backed up in a consistent state that will allow the VM to be spun up from that point in time.  The disadvantage to this is that you may be subject to data lose based on this schedule.

From my testing I was able to get my replica up to date and perform a failover to the replica VM.  You have two options for the failover, a planned failover and an unplanned failover.  The planned failover will ensure consistent date by requiring the primary VM to be turned off.  Once any current replication is complete planned failover will bring the replica VM online and it will act in the capacity of primary VM.  As part of the process reverse replication is established to allow you failback without data loss.  The second option is an unplanned failover.  This failover will allow you to initiate a failover from the replica VM.  If you have enabled recovery point history you can choose which recovery point you would like to recover to.  This can be useful in situations where you have data corruption on your primary VM.  Once the failover is complete the replica VM will act as the primary.  In this case reverse replication must be manually started when the primary site is available.  This will allow you to perform a planned failover back to your primary site when the VMs are back in sync.

My testing has made it clear to me some huge advantages of this new feature.  In my environment I am running my primary VMs on an EMC SAN.  Microsoft has made it possible to perform a replication to a local disk.  This gives you a great deal of flexibility at your DR site.  Windows Server 2012 R2 is making some improvements to Hyper-V replica in terms of more granular replication times and multi site replication.  I hope to explore these features soon and report back with some of my findings.

Agile infrastructure- IT as a service

Private cloud is the name of the game. To be able to quickly adapt to needs of the business we need to clearly identify our service catalog.  The system center suite is the tool set that we have to lay the ground work for our for moving forward with presenting IT as a service.  

There are a number of benefits to viewing IT as a service.  If done correctly this gives you the potential to provide turnkey services that can be deployed in a consistent fashion. This in turn reduces the need for involvement from multiple SMEs and eliminates inconsistencies in deployments.  Also because each deployment is consistent with the last it becomes easy to value the services that you offer.  This gives business leaders some insight into what costs are associated with each request and gives IT the ability to charge back for services provided.  This will allow for a clear view of the value add IT is providing.  

As I mentioned the System Center Suite is what will enable us to achieve this goal.  Through service manager we a defining our service catalog and are making this available to our users.  Through SCCM and SCVMM we are capturing and packaging those services for deployment.  Once those  services are deployed we use SCOM to monitor and alert on loss and degradation of service.  SCORCH is providing us the ability to orchestrate the processes we encounter on a regular basis.  And finally SCDPM provides us the ability to recover our services in the event data loss or corruption occurs.

Once this ground work is laid we will be able to rapidly respond to the needs of the business.  This will also give us the opportunity to develop new offerings and streamline processes through continuous improvement.

EMC World

At EMC World in Las Vegas.  Did a hands on lab for ESI and looked at how the EMC VNX platform can be monitored through SCOM 2012.  ESI also looks like it will integrate well with SCORCH.  Lots of stuff to look at when I get back.

Update: Hyper-V 2012 on Cisco UCS

After running in to some issues while trying to pass the HBA through to the VM.  We have come up with the solution.  I was working on this issue with Cisco TAC and we found that in order to pass that Physical HBA through to the VM we would need to upgrade to the new 2.1 firmware for UCS.  This firmware introduces VM-FEX for Hyper-V 2012.  Here is a link to the Configuration Guide for VM-FEX.  VM-FEX is required to pass the physical HBA through to a VM.  This has been around a while now for VMware and is now available for Hyper-V 2012.  We are currently working on the upgrade to our UCS environment.  I will continue to update on our progress.

Hyper-v 2012 on Cisco UCS

We are working a Windows 2012 Hyper-V cluster. One of the features we are exploring is the ability to use the virtual HBA on the VMs to provide direct access to the SAN and other devices on the fiber channel network. As we continue to work through this I will post our findings.