olly - Fotolia

Scale HC3 hyper-convergence gives firm DR for Hurricane Matthew

A Florida hedge fund firm added a Scale Computing HC3 hyper-converged node at a DR site and replicated data from a three-node HC3 cluster on site to prepare for Hurricane Matthew.

Note: this story was edited from its original version to clarify that Kawa upgraded its disaster recovery plan with Scale rather than implanted DR for the first time.

With Hurricane Matthew bearing down on Florida's east coast, Kawa Capital Management turned to Scale Computing’s HC3 hyper-convergence to protect its data before the storm hit.

The hedge fund only had a few days to put its DR plan in place to secure its business before Matthew hit the Aventura, Fla., headquarters. Kawa already had a three-node Scale Computing HC3 hyper-converged cluster in place, and sought to use it to replace its older disaster recovery set-up.

"The storm was forecast to have a large impact on our location," said Dimas Rosario, director of information technology at Kawa Capital Management. "It was to hit our area on Thursday [Oct. 6] at midnight. And by 3 p.m., our entire building was shut down, with nobody present."

On Monday, Oct. 3, Kawa added a single target Scale HC3 node at internet service provider LightBound LLC's data center in Indianapolis for DR. That complemented the three-node cluster Kawa installed at its Florida headquarters in March.

The sites are connected through an IPsec over Fibre Channel tunnel. Kawa initiated replication Monday night, and then tested the configuration for access and authentication procedures. HC3 DR is enabled through ScaleCare Remote Recovery Service, which provides continuous replication and failover to protect virtual machines and application workloads at a remote site.

Weather reports forecast that Hurricane Matthew was to hit Florida by the end of the week, and it was described as a Category 5 storm.

"By Tuesday evening, we decided to shut down the office on Thursday because the hurricane was forecast to hit our area," Rosario said. "So, from Monday to Thursday midday, we did the change-block replication."

Completing the recovery

The failover from the primary site to the secondary site started at 11:16 a.m. ET Thursday and completed at 11:36 a.m. After the danger passed, Kawa initiated failback at 7:21 a.m. Friday.

The Scale HC3 platform combines storage, servers, virtualization and management in one chassis system.

"It was completed at 7:42 a.m.," Rosario said. "We were back up and running in our production environment in the Aventura office."

The Scale HC3 platform combines storage, servers, virtualization and management in one chassis system. The one-node Scale HC3 recently launched as a cost-effective appliance for DR for SMBs. The single node has the same software stack as the three-node minimum Scale HC3 clusters, including snapshots and asynchronous replication. It delivers up to 6,000 snapshots per virtual machine, with no performance hit.

Customers can replicate from a cluster to a single node, a single node to a cluster, or a single node to the Scale Computing cloud. The single node is priced at $7,500.

"A lot of customers have the budget for a cluster, but they cannot afford a cluster for a second site," said Alan Conboy, of the CTO office at Scale Computing. "The disaster recovery target was wildly out of reach for them. There is no reason a target has to be cluster. We found a lot of folks, regardless of the company size, want another option."

Next Steps

Avoid major outages with proper DR testing

DR monitoring software becomes more comprehensive

Create a plan for cloud disaster recovery

Dig Deeper on Disaster recovery storage