One of our customers had recently opened up a new site at a colocation and wanted to run all non-production out of this new site, keeping production in the existing datacenter for the time being.  This would allow them to leverage the investment in their Disaster Recovery site by running a good chunk of their workloads there and in the event of a disaster, all of the non-prod VMs would be powered off to make room for Production.

Our solution for protecting their production VMs consisted of EMC Recoverpoint appliances, 2 EMC VNX arrays, and VMware Site Recovery Manager (SRM).  The new recoverpoint licensing model (by array, up to 300 TB, for both VNX and CX4 arrays) is fantastic and will make this technology more affordable for a lot of customers.

We decided to use this solution to do a planned migration of the 140 non-prod VMs.  The benefit to this is that the VMs don't have to be offline as you are doing the copy/seeding, and the automation/testing you can do with SRM guarantees that things are going to come up correctly on the other side with no surprises.

After getting all the LUNs replicated with Recoverpoint, SRM installed and configured with the SRAs, and the recovery plan built, it was time to test (which we were able to do in the middle of the day due to how SRM handles tests vs actual failovers on the network).  All went well and we scheduled the actual migration.

We started the migration at 9 am on a Saturday morning, and had 140 VMs back up and running in the DR site (on the new Cisco UCS system, which is definitely the next generation of server technology) by 11:30 am.  No data loss, and they were down for less than 3 hours. Good stuff! 

For more information on datacenter migrations, please contact us at info@rutter-net.com