How to work with multiple pools, but have only one share?

[Force Majeure]

Hello wise ones.
I am building 2 shelves of 24 HDD each for archival storage of video footage. Our company makes videos and we need a place to store the archives over 5 years (and I don't want to use LTOs).

I've been learning TrueNAS over the last year or so and I love it. So far, I have "built" 3 TrueNAS servers, 2 of which are running Nextcloud. So I'm not a pro, but I have figured out what I need to get the results that I want.

This archives that I'm creating is for cold storage. For practical and financial reasons, I can't have a backup (2 copies) of all of this data. But I will try to make it as reliable as possible. Any file lost would not be catastrophic, so it's an acceptable risk.

My original plan was to have the 48 drives (3TB each) split into 4 vdev of 12 drives each (RaidZ2), all part of a single archive pool.
I know that 12 HDD per vdev is a bit much, but I need to maximise available drive space. We generate 2-3 TB of data per month, and I need as much space as I can with the drives that I got...

However, I am rethinking this plan because I don't like the idea of having a pool spreaded over multiple disk shelves. I'm also aware that loosing a single vdev would loose me the whole archives.

Therefore, my new plan is to have 4 pools of 12 drives each (RaidZ2).

Here is my question: is there a way of combining the pools into a single share, so that the users only see one share folder; and having the system automatically take care of spreading the data evenly over the 4 pools?
I would like not to have to manually move data from one pool to another when one of them is too full. And I would like to be able to search all 4 pools at the same time if I'm looking for a specific folder.
And I want to be able to add more pool in the future, as needed.

I've done some searching on Google, but I can't find any leads.

Thanks

 

[jgreco]

Here is my question: is there a way of combining the pools into a single share, so that the users only see one share folder; and having the system automatically take care of spreading the data evenly over the 4 pools?

No, not really. If you have multiple systems, Scale "supports" (for some value of that word) treating individual nodes as parts of a Gluster filesystem, with resultant performance penalty and related guinea pig issues. You're probably better off treating it as a single pool.