1

Is there a way to determine the zfs dedup ratio on a per folder basis?

I realize the potential difficulties in computing this as every block would need to be analyzed, but can such a method be scripted?

Dan Buhler
  • 486
  • 5
  • 9
  • 2
    dedup is a property of file systems, not folders. Maybe you are asking "how to find whether folder X contains duplicate information/blocks"? What is your main goal, i.e. what will you do if I give you magic script that will say "folder X in file system Z is deduped 5x"? – aaaaa says reinstate Monica Mar 26 '18 at 03:41
  • It's a Veeam backup repository and I'd like to be able to get an accurate figure of each backup job's data. Also, I'd like to be able to check if one job is suddenly creating data that does not dedup as we will generally get 5-10x dedup on our data. If this isn't happening, we need to know to fix it. And yes, it is definitely a property of file systems, but on other systems like Storage Spaces, you can calculate it on a per file/folder basis even though it's very resource intensive. – Dan Buhler Mar 26 '18 at 16:45
  • There is another consideration: ZFS snapshots are stored as "diffs" so deduplicaiton is kinda built-in for the single folder. When we backup shapshots, we don't transfer same data every day to another server, only changes. If you have new data backed up that need to be deduped, i am not sure how that works. – aaaaa says reinstate Monica Mar 26 '18 at 18:59

0 Answers0