Available Diskspace & snapshots

[fuznutz04]

This morning, I logged into Freenas and found that I now have almost 16 TB of free space. Last week before the weekend, I only had about 7.5 TB. I started to think the worst and thought that something was deleted accidentally. Then I remembered that I changed the snapshot settings last week. I changed the retention from 2 weeks, to 1 week and reduced the snapshots to only 1 per week. Snapshots run at 9 AM, once a week, which is exactly when my storage pool changed (according the partition report.) From what I can tell, changing the snapshot is what caused this big decrease in available space. This server is purely a backup server, with terabytes of changes daily.

This whole thing got me thinking about snapshots in general for my environment. This server is used for 2 things: A VEEAM backup repository, and storage for current and historical SQL backups (run from SQL management jobs) Since VEEAM is setup to have its own history (2 weeks is all that we require), is it really necessary to have FreeNas also perform snapshots, or is this just a waste of space?

Should I split the pool into 2 data sets and only run snapshots on the SQL DBs?

Any insight and discussion would be helpful.
Thanks!

 

[DrKK]

I don't know much about VEEAM, but if it is as you say, then everything you're saying sounds sensible. The granularity for a snapshot is a ZFS dataset, so, yes, segregating that-which-is-to-be-snapped from that-which-is-not by dataset is inherently reasonable.

 

[fuznutz04]

I don't know much about VEEAM, but if it is as you say, then everything you're saying sounds sensible. The granularity for a snapshot is a ZFS dataset, so, yes, segregating that-which-is-to-be-snapped from that-which-is-not by dataset is inherently reasonable.

Sounds good. I only have 1 dataset now, so I'll create another one and enable snaps on it, and disable snaps on the other. Thanks for the input!