Snapshots have a visibility problem

There’s a fantastic bakery in my neighborhood where everything is terrific. Whether you’re getting bread or cakes or cookies, they are the best in town. Unfortunately, there’s one problem.

The problem is that nothing is labeled. When you gaze into the glass cases containing all the goodies you have to guess what you’re looking at. Does that cake have some kind of filling? Is it vanilla frosting or maybe hazelnut instead? What’s under that icing? Chocolate? Carrot cake? You just can’t tell.

So what you have to do is ask and point. “What about that one? No, not that one, the one behind it… yes, that’s it. What does that have in it?”  And so on. It’s maddeningly inefficient. How much simpler it would be if they would just label things so I would know what I was asking for!  What the bakery needs is better visibility.

Which brings us to one of the major drawbacks of using snapshots for data protection: limited visibility.

Snapshots are the best way we have of capturing lots of data quickly and efficiently. But when it comes time to finding specific data within the snapshots, what do you do?  Mostly you guess.

Let’s go back a few steps and take a look at the snapshot process. It begins on primary storage, where disk array software captures snapshots of live production data. Sometimes this is file-based data if you are snapping an NFS or CIFS share, other times you snap LUN-based data, which is even more opaque. Since keeping lots of snapshots on primary storage is expensive – in terms of high-cost disk as well as performance impact – most IT shops keep only a few days of snaps on primary.

To continue reading this article register now

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon