en

Data Protection for Nutanix Files

Nutanix users have a very convenient way to create file storage for their applications — it is called Nutanix Files. In this short blog I’ll introduce how vProtect can help you to backup these file systems

With vProtect 4, we introduced Storage Providers as a new source type, that in general is to focus on storage data protection (instead of VM-centric or App-centric approach).

We started with support for:

  • Ceph RBD
  • plain file systems
  • Nutanix Files (AFS)

The last one is in general the extension for regular file systems data protection mechanism.

How does it work?

Nutanix Files allow users to create shares either via SMB or NFS protocols, which later are mounted and used by the clients:

From the vProtect UI – every File Server is a separate Storage Provider, which owns set of “Storage instances” — in this case file systems. When you delete a file system from Nutanix console, you still can restore backups from the vProtect Console:

vProtect talks directly with Nutanix APIs to grab the list of the file systems and to grap list of changes between full and incremental backups.

The overall process looks like this:

  • vProtect fetches information about the file system from Nutanix API
  • vProtect creates snapshot to have a consistent point-in-time backup
  • vProtect mounts share on the Node, so the data is locally accessible on the file-system level
  • Standard backup mechanism for file systems is used — to copy all of the files and directories from the share
  • Once it is complete, share is unmounted
  • Last snapshot is kept for future incremental backups, any previously created vProtect snapshots are removed
  • Data is uploaded to the target backup destination

Incremental backups are created using Nutanix CFT (Changed-File Tracking) API, to fetch information about files that have changed since last backup. This is a different approach than in plain file system backup, where vProtect would scan file system for changes. Obviously – querying the API is significantly faster, which is especially helpful, when not to many changes occur in the file system. Incremental backups using this technique reduce time on huge file system — this wouldn’t be possible with plain file system mounted over NFS/SMB with regular file-system scanning.

Restore process works exactly in the opposite direction:

  • vProtect creates a new file system on AFS
  • vProtect mounts share on the Node, so the data is locally accessible on the file-system level
  • Merged backup chain is restored to the new share
  • Once it is complete, share is unmounted

Backup and restore process doesn’t require any agent installation in the Nutanix itself — all of the communication and data transfer can be done over LAN using access to the shares and API calls.

vProtect also allows you to restore individual files from the backup that have been created — without the need to import data to the file server.

Scalability

With large file systems it would be hard for any backup solution to create backup (especially with thousands of files hosted by many file servers) within reasonable time. With vProtect you can have multiple nodes, each one protecting its own file server, so you can spread the workload on multiple nodes.

Wrap up

vProtect 4 is a complete data protection solution for Nutanix platform. It allows not only to protect VMs running on Nutanix AHV hypervisor but also file systems in Nutanix Files. With scalability using multiple nodes it is easy to protect even very big Nutanix Files deployments. Last but not least, all of these backups can be stored in many supported backup destinations.

When considering backup for Nutanix it is always good to ask:

  • if the solution supports Nutanix AHV – both regular backups and snapshot management
  • if the solution is able to protect Nutanix Files with CFT
  • if the solution is able to restore individual files from backups

For all of them vProtect answers yes.

Marcin Kubacki

text written by:

Marcin Kubacki, CSA at Storware