danb35
Hall of Famer
- Joined
- Aug 16, 2011
- Messages
- 15,504
I think there's a fundamental misunderstanding here. A pool ("volume" in FreeNAS terminology) is a filesystem, and it is in fact a dataset. While it's possible to make sub-datasets of your pool, there's no need to do so. It would be common to create separate datasets for each share, if you had multiple shares set up, for example, but I can't think of any reason or benefit to creating a single dataset within your pool. That doesn't address the question of why your dataset capacity doesn't match your pool capacity, but it's something you should consider.thanks, pirateghost for your answer.
I have no problem with 19TB volume, but I want 19TB filesystem too. As far as I understand, in zfs world dataset it's filesystem...
Second, your pool configuration is just bizarre. You have two five-disk RAIDZ2 vdevs combined with a two-disk mirror. I don't know how or why you did that, but that's what you have. This will give you the net capacity of 7 of your disks, which would be 14 TB, or 12.something TiB. This is pretty close to what FreeNAS is reporting as the capacity of your dataset. The question is why it's reporting the capacity of your pool as 19.something.
A single RAIDZ2 vdev of 12 x 2 TB disks (which is wider than recommended for a single vdev, but that's a separate issue) should have ~ 17.5 TB of capacity. You could build such a pool by detaching your current pool, checking the box for "mark disks as new" (which will destroy your data), and then building a new pool from the GUI following http://doc.freenas.org/9.3/freenas_storage.html#volume-manager. Add all your disks, set the drop-down menu to RAIDZ2, and pull the slider all the way to the right. However, your installation still thinks you have another pool attached called "pool", likely due to your mucking about with the CLI. If you can remove that entry from the GUI, it'd be a good idea to do so. If not, your config database is likely borked, and it'd be a really good idea to start over from scratch.
Last edited: