Reporting: Datasets showing misleading/incorrect sizes

[underpickled]

I can post a screenshot later, but here's what I'm seeing...
Basically, I've got a dataset called "Media" in my main volume and another called "Private". My main volume has a total capacity of 10.7 TB, Media is taking up about 2.6TB, and Private is taking up maybe 400 GB, give or take.

In reporting, the main volume shows a total space of 7.6TB, and all of it is free space. Media shows a total space of 10.7TB, but with 2.6TB used. Private shows the same thing as the main volume... the space is taken up in a child dataset of Private.

Basically, parent datasets' space usage do not include their children. This results in the main volume reporting an incorrect size, while the dataset using the space indicates the true full size.

 

[dlavigne]

This is a known bug. That collectd reporting module depends on the output of 'df' which makes incorrect calculations for the 'free' value ZFS filesystems (it assumes total_size-used=free which is not true). It should be using 'zfs list' for ZFS filesystems.

 

[underpickled]

This is a known bug. That collectd reporting module depends on the output of 'df' which makes incorrect calculations for the 'free' value ZFS filesystems (it assumes total_size-used=free which is not true). It should be using 'zfs list' for ZFS filesystems.

Thanks for the clarification. Wanted to make sure I didn't accidentally screw something up.