We help IT Professionals succeed at work.

Offsite Replication Solution

I currently have two VMWare servers that host 5 virtual machines running Windows Server 2012. One server is the main, and the other has replicas. I would like to have an additional offsite replicated server for 3 of the 5 servers. I am implementing new servers so I am open to using alternate software on the host. But this decision will be based on the ability to achieve reliable offsite replication. Is this something a new version of VMWare can handle? Or is there a better software solution, such as Hyper-V that would be a better option.

Thanks in advance!
Comment
Watch Question

Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
Depending upon your Shared Storage Solution, or Backup Solution, many now offer offsite to Amazon AWS Storage, which is very cost effective.

Shortly VMware will also be offering Amazon AWS additional services.

Would you like to store copies of backups, or BRING ONLINE Virtual Machines, should you have a failure ?
Bill BachPresident and Btrieve Guru
CERTIFIED EXPERT

Commented:
Do you want host-based or guest-based replication?  I have had great success with Double-Take Availability, which is very inexpensive for a virtual machine.  It runs inside the guest OS, and you can opt to replicate the entire VM, or you can opt to replicate just the data within the VM, leaving the OS and other services separate.  (This buys you protection from a bad OS patch or malware that takes out the OS.)  By only replicating the changed data (and optionally enabling high compression rates and bandwidth throttling), this works very well for large data sets and WAN links.

For servers where immediate access to the backup server is not required (i.e. you have time to do a restore process), the Double-Take DR solution works similarly.  It replicates changing data just like Double-Take Availability, but it sends it to a repository server for storage.  In the event the production box fails, you set up a Restore job from the repository to a newly-created VM, and when it is done restoring, you boot it up, with all of your data intact up to the point of the failure.  Price-wise, Double-Take DR is about 1/4th the cost, so this makes a lot of sense if you already have local replication as protection for system failure, and if you only need some extra (off-site) peace of mind.

Author

Commented:
Andrew,

I would like to bring them online in the event of a failure.

Bill,

I want to keep the current configuration (having onsite backup and replication) for local fail-over. And I believe I would prefer to host the solution myself, but I am open to looking into hosted options. But I am not very familiar with the hosted solution and how it functions (we host SQL, and IIS on site ourselves that is accessed from the internet as well as local workstations)

And activating the replicas would need to be instantaneous (or as close to that as possible).

Dan
VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017
Commented:
I would wait for VMware on Amazon AWS, you then don't have to do anything, other than create an Amazon AWS account, and pay for usage! (e.g. when you bring them online!)

https://www.vmware.com/cloud-services/vmware-cloud-aws.html

or you can use Hotlink today

http://www.hotlink.com/
Bill BachPresident and Btrieve Guru
CERTIFIED EXPERT
Commented:
I think you misunderstood my question.  A "host-based" replication solution works at the hypervisor level, where you have nothing to place inside the protected OS itself.  A "guest-based" solution works inside each VM, interacting directly with the local OS itself to protect the environment.

Double-Take can handle both your (local (on-site) and remote (off-site) replication needs, or you can use it for just the off-site needs -- whichever makes more sense for you.  To minimize the restore time, Double-Take Availability would be the best option, as it can keep the backup machines either "hot" -- ready to set up the network connection, start services, allow connections, and go -- or you can make the machines "boot-ready" -- i.e. when a failure occurs, you would tell DT to boot the new VM's and they would be ready within a few minutes at most.

There are even multiple licensing options, which may depend on how many machines you want to protect.  You can cut costs with a Virtual Recovery Appliance (VRA), which is a single DT machine in the target environment that receives data from each of your source machines and keeps the virtual hard disks for the protected machines mounted locally.  In the event of a fail-over, the VRA contacts vSphere to dismount the volumes from itself, mount the volumes back to the target VM instance, and then boot the VM.  You get 2-minute response time to a live copy of production with a cost savings to boot.

Author

Commented:
Based on your description I believe I would prefer a host based solution. I am going to do some testing with HotLink and Double Take and see if they work for me. I will accept both your answers as solutions and post back after I find a solution that works.

Author

Commented:
I will be looking into these solutions and will post back later.

Explore More ContentExplore courses, solutions, and other research materials related to this topic.