«

»

Feb 22 2014

Disaster Recovery – Failover and Failback with Nutanix

Advertisement

Nutanix has built-in VM-Centric, multi-site, bi-directional, multi-topology disaster recovery and replication engine that also supports VMware Site Recovery Manager and Run Book Automation tools. The replication utilizes incremental and fine-grained byte-level data transfers with intelligent data compression, eliminating network and storage resource throttle.

Administrators may completely failover entire deployments to a secondary datacenter and later on failback with all data created in the secondary datacenter already replicated back to the primary datacenter. As of today the replication is asynchronous for up to an hour RPO/RTO as the most granular recovery point, but this is likely to change in the near future in favor of a more granular approach.

 
Protection Domain – A protection domain is a group of VMs to be backed up locally on a cluster or replicated on the same schedule to one or more local or remote clusters. A protection domain on a cluster is in one of two modes:

Active: Manages live VMs; makes, replicates, and expires snapshots.
Inactive: Receives snapshots from a remote cluster.

Consistency Group – A consistency group is a subset of VMs in a protection domain. All VMs within a consistency group for that protection domain will be snapshotted in a crash-consistent manner. For all VMs in a consistency group, a snapshot creates one snapshot for all VMs in the group.

 

I recorded the video below in my lab that demonstrates the failover and failback of a Citrix XenDesktop machine with changes being applied to the VM while the VM is in the secondary site. It might just be me, but I have never experienced or seen a multi-site DR solution as simple as what Nutanix has built into their product. Moving forward I will also produce more technical articles that depicts how Nutanix disaster recovery works under the covers.

 

(Highly Recommended to watch in FullScreen 1080P)

 

In my article Nutanix Automation, Policies and the SDDC I discussed how Nutanix exposes all features and functions via REST API, enabling programmatic access to datacenter services within mainstream enterprises, and in my last article I demonstrated How to Establish PowerShell Connection to Nutanix and Execute your 1st Query.

The script below demonstrate how to trigger disaster recovery for a protection domain using the REST API via PowerShell. The source protection domain is in the URI and the remote site (destination) is in the Body of the REST request.
 

$Uri = "https://10.20.18.10:9440/PrismGateway/services/rest/v1/protection_domains/bizdev01/migrate"
$username = "username"
$password = "password"
$Header = @{"Authorization" = "Basic "+[System.Convert]::ToBase64String([System.Text.Encoding]::UTF8.GetBytes($username+":"+$password))}
$Body = "POC07"
Invoke-RestMethod -Method Post -Uri $Uri -Headers $Header -Body $Body

 

Note: Stretched VLAN is required for seamless DR and fallback capabilities. Optionally, VMware Site Recovery Manager (SRM) or other Run Book Automation tool can be used to manage IP addresses and start sequence in each DR datacenter.

 

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Similar Posts:

Permanent link to this article: http://myvirtualcloud.net/?p=5945

6 comments

1 ping

Skip to comment form

  1. Josh Sinclair

    We actually have used the DR to perform a couple of datacenter migrations for some web apps. With protection domains you can actually stage the migration days in advance to give yourself enough time to transfer all of the data. Then you just shutdown the VMs, either force a snapshot or wait for the scheduled one to complete and then activate the protection domain at the remote site. Minutes of downtime instead of staying up all night trying to migrate everything during the maintenance window.

    The best part is that all of that is accomplished with 3 commands in the CLI. The caveat is that it only supports 25 VMs per protection domain and it is a separate licensed feature, even though it is baked into the NOS.

  2. Greg Lenczewski

    Hi Andre,

    This is interesting. Is this a new feature that has been added to the Nutanix?

    I’m kind of curious if Nutanix would support stretched cluster across two data centers on a stretched VLAN? So the end result would be that there are two separate geographical locations, however vSphere would see them as a local cluster.

  3. Andre Leibovici

    Hi Greg,

    Today stretched cluster is not supported. However, Nutanix will soon be able to manage and monitor multiple clusters via a single interface and also be able to move workloads between clusters. I can’t tell you much more at this point in time. Please note that all that may not come together on the same release.

  4. Greg Lenczewski

    Thanks Andre. I appreciate your feedback.

  5. Andre Leibovici

    Hey Josh,

    Thanks for your comments.

    Nutanix does have a limitation of 25 VMs per PD. This restriction was put as QA has tested/qualified max 25 VMs per PD, but the back-end was built to handle thousands of VMs per PD.

    As soon this number changes I will post an update here.

  6. Josh Sinclair

    We were able to talk this week with engineering and they cleared up some confusion regarding Protection Domains for us. We were previously led to believe that we could only have 1 PD per container, however we were told this week that there is no limit to the number of PDs per container.

  1. Nutanix has never lost data… and here is why! » myvirtualcloud.net

    […] here. I am going to discuss backup and DR in a future article, but in the meantime you may watch this failover and failback video I recorded a while […]

Leave a Reply