Defrag, snapshots, volume usage

Snapshots use CoW (Copy-on-Write). Even if CoW is disabled on a share, CoW is still used at the point in time a snapshot is taken. A defrag breaks the CoW link between the current data and its snapshots.

Our engineers are looking into the snapshot space indicator problem.

1. Yes. The more you change, the more increased usage you would see.

2. Defrags are good. If your volume is getting quite full it would be a good idea to free up some space before running a defrag.

3. Snapshots will still work but they will consume more space and thus make less efficient use of space.

4. Yes, they are just as good practice to run with snapshots as without. A balance moves data and metadata around in order to empty chunks so they can be returned to unallocated space. Again you should keep volume usage to reasonable levels. If volume usage is very high then especially if a balance has not been run for quite some time it's possible the volume will become fully allocated and amongst other problems a balance may fail (though on 6.3.x this is less likely to happen).

Backups are important. No important data should be stored on just the one device.

---

vandermerwe wrote:

In this thread, 

https://community.netgear.com/t5/Using-your-ReadyNAS/Performance-Admin-Page-Access-Does-available-disk-space-affect/m-p/980153#M93263

one of the posts mentions that performing defrags in the setting of snapshots leads to an increase in volume usage.  

I suppose if the snapshot space indicator in 6.4.x works properly in the final release we will be able to see just how much more space is used.

I currently I have 4 defrags scheduled each year and I use daily snapshots on 4 shares, data in these shares doesn't change much.

 

Questions:

 

How much more space will be used up - depends on how much data changes I suppose?

  

 

---

It's a combination of how much data churn there is, and how much fragmentation that churn creates.  The link is only broken when the main file (a) is fragmented and (b) has older versions stored in the snapshots. 

For instance, in my own case the OS6 systems are used to back up my Pro using rsync.  When files change on the pro, rsync copies the full files to the OS6 shares.  Churn might be high on some shares, but fragmentation won't occur as long as I keep reasonable free space on volume.  So defrags don't increase my snapshot space usage (and in fact aren't needed much, though I do them anyway).  

If on the other hand, you had an SQL database that you are frequently making small changes to...  CoW might be quite efficient, keeping the snapshot usage low.  That database would also become very fragmented.  In that scenario, a defrag would result in a pretty big increase in the snapshot size.