Can two Nimble arrays act as backup and DR?

[hendersond]

Traditionally Disaster Recovery and Backup have meant two different things. Disaster Recovery is the ability to quickly recover from a big disaster in a main data center by turning data in a DR site into production data quickly. This usually means some type of replication from the data center to a DR site. Backup usually meant backing up data using some type of software solution to control the backup and some type of target (other than the array in the DR). The backup had some type of retention with the ability to do restores on the volume, VM, or file level.

Current setup

1. Two arrays from same company, one in data center and on in DR site with replication running between them so my DR strategy is covered
2. Commvault software running that backs up data in data center to Exagrid box in DR site so my backup strategy is covered

Lots of moving parts and costs associated with Commvault and Exagrid

If I go to this setup

1. Two Nimble arrays, one in my main data center and one in my DR site
2. Snapshots are happening on the array in the data center with snapshot set to go back 90 days
3. Replication setup from my data center to DR site

Can this handle Disaster Recovery AND Backup? In other words, if need be I can:

1. Turn the data in the DR site into production data in case of a fire/flood in my main data center
2. Do volume level, VM level, or file level restores

In the end can I ditch Commvault and the Exagrid box by purchasing Nimble?

[dsandler71 · Accepted solution]

Hi David.

The biggest question here, IMO, is do you need to have a catalogue of your backups. Currently, CommVault allows you to search for any given item you wish to recover. If you are just relying on Nimble snapshots, there is no native catalogue/search capability. That being said, if you don't need the catalogue, then we can 100% accomplish your goals. In fact, we have a large (and growing) number of customers that do just what you've described. Once you're on Nimble in both sites, you can have the same 90 days of retention (or more if needed). One of the nice things with Nimble is you don't need 90 days at each site if you don't want to. Since you can set independent source and target retention policies, you can keep much shorter retention on one side. And if the time ever comes (hopefully not), you can make the DR array primary for all or a part of your workloads. This can even be handled through our Handover capability for planned transitions.

Hope this helps!

[hendersond]

This does help. We have been a Commvault customer for a long time and other than searching an email archive, I have never used the search capability. Most of our restores the person knows a date we can go back to. I am assuming if we went with Nimble we would just go back to a snapshot on that date, mount the snapshot and start drilling down to find the file. How easy is it to do this?

Moving away from Commvault and our Exagrid box for backup makes me excited in one way (less moving parts and less cost) but also makes me a bit nervous as well. There are hundreds of articles that talk about the difference between backup and DR and my hope is two Nimble arrays with snapshots can handle both sides of this

[dsandler71]

Your best course of action would to find the specific snapshot you want to recover and create a zero-copy-clone. This will create a new volume instance that looks like the original volume did at time of snapshot. The benefit is that it is its own entity and can be resigniatured when mounting up to a host for recovery.

[hendersond]

And once I mount this up to a host, can I drill down and do file level recovery? Can I recover an entire VM?

[dsandler71]

Yes. You'll have a full view of the filesystem. You can browse the datastore and recover the VM or use something like UFSExplorer to get to the file level.