I don't Understand Snapshot size

[mcarrara]

I am new to setting up a FreeNas system. Now that things are working with sharing and my users have access to files, I am ready to run snapshots then store them offsite.

Here is what I don't understand. I did a manual snapshot of a dataset with 236mb of data. This was the first snapshot and it was 1.42 mb. I took two snapshots from other datasets manually. One dataset was 550 Gb and the snapshot is 504 kb. The last one was 1.1 tb and the snapshot was 2.4 mb. I get this data from the snapshot tab of the web interface. I am using version 11.2. Am I not understanding what the interface is showing or did I do something wrong? Shouldn't the first snapshot be about the size of the dataset?

Mark

 

[fracai]

The reported size of the snapshot is only the data that is unique to that snapshot. If you take a snapshot and immediately look at the size, you're going to see a number very close to zero, or maybe even actually zero. This is because, depending on how your dataset is being used (static storage, constant logging, etc.), there may be nothing changed from the snapshot to the current dataset state. If you then delete some files from the dataset and check the snapshot size again, you'll see that the snapshot has appeared to grow in size. This is because the deleted files are now only referenced by the snapshot. It's like a Venn diagram and the snapshot is a circle that only reports what isn't inside any other circles.

This can get even more tricky when you get multiple snapshots. They'll all only report how much space is unique to each snapshot. This can lead to situations where you delete very large amounts of data, but the snapshot sizes don't change because the data is still referenced by multiple snapshots.