Tech/Engineering

The complexities of managing typical backup systems

A recent conversation with a good friend of mine reminded me of just how hard it can be to administer a typical on-premises backup environment. He has had no end of problems with it, and worries about the integrity of his backups as a result.

I won’t say which backup product he is using, but it is one of the leading on premises backup solutions that has deduplication integrated into the product. My friend opened the conversation by saying, “You know, when we switched from Product X to Product Y, I was leery of their claims about how good their dedupe was. But now that I’ve had a lot of time with it, I can say I truly hate it.” Obviously my friend was going for a laugh, but he continued.

Understanding the pain of on-premises backup solutions

He explained that the dedupe database was continually getting corrupted. The vendor told him this was no problem, because their dedupe data is self-describing and the database can easily be rebuilt. You just need to shut down all backups to do so! My friend said he has needed to do this several times, and each time he does it, he is without backups for 2-3 days! It is correct that the database rebuilds itself, but in the meantime, he has no backups or restores. While he agrees the rebuild seems like a nice feature, it’s one he wishes he didn’t need so much.

My favorite part of his story was when he described his problems to tech support, they said they thought that maybe he needed faster storage for his dedupe database. (He’s using NVME flash, so I guess they think he needs Intel Optane.) I personally found that idea idiotic. The dedupe catalog should not get corrupted even if he was using a 5000 RPM SATA disk. It would probably be incredibly slow, but why would slower disk result in a corrupted database?

Some familiar with his backup product think that he might just be trying to do too much with one backup/dedupe catalog, as he is doing several hundred terabytes. (This matches what I’ve heard about many of our competitors’ products; their dedupe stops working well after reaching a certain size.) Their suggestion is to split the backup environment into two backup systems, which means more cost and more management.

My friend is also using a leading storage vendor for snapshots and backup data, and he started talking about the management of these storage arrays as well. He’s continually having to query each volume, trying to figure out how much storage is actually used, and how much is being consumed by each activity. Then once he realizes a volume is running out of capacity, the hoops he has to jump through to buy and add more capacity sound simply horrendous. Part of this is the typical capital procurement process of a large company, and the other part is the actual physical management of the storage array and its volumes.

It’s been a few years since I had to do real storage management in a data center, and I thought things might have gotten better since then. But that sure didn’t seem to be the case. I listened to my friend’s woes while shaking my head – all of this work just to get some backups done.

The advantage of using SaaS data protection

One problem of coming to a vendor of one product is you don’t get to spend a lot of time around those who use other products. Our customers don’t have any of these issues, so I simply forgot how hard it can be to get some backups done. Our dedupe catalog is incredibly scalable and backs up customers larger than 10 PB all to the same dedupe index. I recently learned of a customer that – in addition to backing up many large data centers – recently backed up a 1 PB Isilon to our cloud solution.

Using cloud-based object storage as the backup repository also provides unlimited storage capacity with zero management overhead. It grows as you need it, and cleans itself up if you prune some of your backups. No volumes to worry about, no disks to buy – just unlimited capacity that you pay for by the terabyte by the month.

As I often say, the direct-to-cloud data protection-as-a-service (DPaaS) model Druva employs doesn’t work for everyone. But if it could work for you, why would you want to be forced to manage a backup system like my friend does? Why not just use a service and get rid of all that hard work? Backups don’t need to be so hard.

Want to learn more?

The Druva Cloud Platform, a SaaS data protection and management platform, delivers a secure, scalable and easy to manage solution while saving our customers up to 50% on their TCO.

Druva and sfApex deliver the best of SaaS data protection and management for Salesforce | Learn how sticky-promo-icon-carrot Druva and sfApex | Learn more sticky-promo-icon-carrot