Some very exciting news came out of Ignite on May 4th (insert star wars joke here). And it didn’t get much hype. For the first time in a LONG time I got super excited over technology. The Azure Site Recovery team and the SharePoint team announced that ASR would be formally supported with SharePoint deployments. This was an effort months in the making with a lot of testing and dotting of I’s. From both teams.
I feel like it is the first time you can quite literally have a rock solid DR plan that is: Cheap, Fast and Easy. Something up until recently would have only allowed you to choose two of those options. ASR is in a nut shell (although WAY COOLER than..) is VM replication to Azure in real time. Awesome enough it also does Physical to Azure.
The jist of it is this. VMWare, HyperV, VCenter, Physical get a synchronization job running on the VM/Physical host the job is replicated up to Azure where then Azure sucks in those bits and drops them into Azure Storage and nests them into a fail over job (Recovery Plan) that allows you fail over to Azure with the click of a button, as well as failing BACK. Literally 1 click (not including confirmation clicks). You can also test your fail over with out ever actually failing off of your production environment. Azure will replicate your VMs on the fly and simulate a real life fail-over that is exactly the same as if you went from on-prem to Azure in a real disaster type situation. SUPER FREAKING COOL! No more freaking out about either accidentally bringing down your real environment OR stressing over if your fail-over plan will actually work in real life. Now you actually know!
And its insanely cheap insurance. $54 dollars per instance, per month with a bunch of built in incentives for EA customers. I spent more on that on a cab from ignite back to the airport! And the only time you pay a cost for the VMs outside of storange and data is when the fail-over switch is actually triggered. So you no longer have to keep servers running for a heart beat to make sure data is flowing and incurring the costs of those VMs in Azure.
Watching the guys do the demos for this a few months back while in private tap and getting my hands on it to play around left me craving more! The teams did just that, I spoke about physical to azure as well. Currently if you have a physical box in your SharePoint farm, lets say an Index server. When you fail that physical over to Azure, Azure converts that to a VHD and runs the VM in that state. When you need to fail back however, the fail back will push the VHD version of that physical instance back to your VMHost on-prem. Now under initial thought, that kinda stinks but if you think about the technical challenge of creating a physical box, kernel level, drivers, patches etc etc from a virtualized file is pretty intense. So when you find that your butt was majorly saved failing over and keeping what Gartner (i think it was them) estimated as .5-1.5 MILLION dollars per hour in downtime lost for major corporations adding a physical machine back into your SharePoint farm and offloading the indexing (or whatever) capabilities from the VM that was failed back, to that newly (or re-inserted) physical machine is pretty tolerable.
All that said, the Physical to Azure part is currently in private beta so you can request it, hopefully you’ll get to play around with that soon enough.
Check out the Slide deck I’ve added from my talk at the SharePoint Evolutions Conference. I walked through how to do this and what planning is involved. I’ve updated the deck to reflect that it is now supported vs when i did the talk a week prior, we knew it was coming but couldn’t talk much to it. A video of the actual talk as well as the video from the VM to Azure demo is below as well.
Excited about backups.. Never thought I’d see the day.
Video and Session walk through (remember, at the end we talk about its WILL be supported. This was given prior to ignite. Azure Site Recovery IS NOW supported. Check it out over at SharePoint Evolution Conference site.
Just the video on setting up a Hyper-V to Azure fail over:
Updated presentation deck: