Please enable javascript, or click here to visit my ecommerce web site powered by Shopify.

Community Forum > Savings from compression

I created a couple of volumes in a Storage Pool that had compression enabled (by default). One volume (A) has no data written to it while the other (B) has ~180MB of data in it. When I right click the volumes in the GUI, the empty volume, A, shows a compression ratio of 22.28 while volume B shows a compression ratio of 22.24

Does this mean that the 180MB of data in volume B has been compressed to a size of 180/22.24 = 8.094MB?

I am looking for any doc that describes how to measure the savings from compression (and deduplication). I looked at the Admin guide and the Getting Started guide; neither of them helped.

Regards

March 10, 2014 | Registered CommenterGopi Rajagopal

Hi Gopi,
Yes, for highly compressible data you'll see high ratios like that. In practice though we generally see ratios more like 1.5x to 2x compression for workloads like virtualization. QuantaStor uses the standard lzjb compression algorithm but ZFS supports a variety of compression options which can be set at the command line. Going forward we'll be adding options like None, Low, Medium, High to the Create Storage Pool dialog so that you can configure higher levels of compression for archive applications when you create the storage pool without having to use the zfs command like tool.
Ben Rockwood has an interesting post on his blog here that goes into more detail on the compression technology and where the compressionratio property is found in a ZFS filesystem. We collect that information automatically so you'll see it change around as you write more information to the pool, shares and volumes.
Best,
-Steve

March 10, 2014 | Registered CommenterSteve

Hi Steve,

Thanks for the information. A few questions -

Is there a way to obtain the space savings on a per volume basis?
Is there a tool that can provide an estimate of the savings (similar to the Space Savings Estimation Tool from NetApp or the compresstimator from IBM)
Is the compression inline or post-process? Is it possible to choose between the two?
Is there any data on what kinds of datasets benefit from compression?

If there is a guide, please point me to it.

-Gopi

March 11, 2014 | Unregistered CommenterGopi

Hi Gopi,
Yes, in the Storage Volume Properties section you will find the 'Compression Ratio' for the volume selected in the tree view. Similarly for Network Shares you will find the compression ratio in the properties page. We should add this as a column to the table view in the center.

>Is there a tool that can provide an estimate of the savings?

Just the compression ratio information. Is that what you mean?

>Is the compression inline or post-process? Is it possible to choose between the two?

Inline compression only.. no post process.

>Is there any data on what kinds of datasets benefit from compression?

Virtual machines and most general purpose workloads benefit from compression. Data that is already compressed like jpg files and mpeg video files don't compress well if at all. So it really depends on whether or not the content is already compressed or not.

Best,
-Steve

March 11, 2014 | Registered CommenterSteve