Aug
24

Scale Out Disaster Recovery with Nutanix and @VMware SRM

VMware’s Site Recovery Manager (SRM) a disaster recovery management for virtual applications.

VMware host based replication uses VMware snapshots and replicates the snapshots from host to host. VMware snapshots suffer from degraded performance on reads to snapshots with long snapshot hierarchy, and collapsing delta files is extremely slow.

The Nutanix SRA leverages native remote replication, which provides faster and more efficient scale-out replication of Nutanix snapshots.

To use the Nutanix SRA, you must first configure the Nutanix cluster then configure the SRM servers.

Nutanix SRM Architecture

SRA

Simple View of the Nutanix SRM Architecture

Both the protected site and the recovery site have a vCenter server that manages protected VMs and an the SRM server. The Nutanix SRA is a set of scripts that must be installed on the SRM servers. The SRA uses the Prism REST API to communicate with Nutanix Operating System(NOS). NOS talks to Cerebros which runs on each host in a scale out fashion. Since recents changes are in flash and all the controller VM’s replicate the data it helps to eliminate bottlenecks in the infrastructure.

vStores

A vStore is a separate mount point within a container which has its own NFS namespace. This namespace maps to a Nutanix protection domain. Each vStore is exported as a device through the Nutanix SRA. Since Containers in Nutanix are just an away of applying a policy to a group of VM’s, it’s business as usual. You will have to move your VM’s into the new container associated with protection domain. SRM does have some baggage from the “LUN” world. Nutanix’s native replication does not have this requirement.

Come by the DR booth at VMworld 2013 where and I will fill you in on all of the Techno Wonky Bits.

-DL

Speak Your Mind

*