Dataset layout recommendations

Status
Not open for further replies.

agartha79

Cadet
Joined
Aug 23, 2017
Messages
8
Hi all, first time forum user here!

I've just assembled my new nasbox and need to setup a proper :) and flexible dataset layout.

What do you guys think of the following layout? Any suggestions?
Code:
storage1 (encrypted, 2x4TB WD-Gold)
	storage1
		backup
			<some host's name>
		home
			<some user>
				archive  (added cron job that changes owner to root on all files at given time a day so that they become read-only)
					media
					projects
				documents
		jails
		repos
			main-gitrepo (using nfs)
		vm
			gitlab1 (zvol, debian 8 running docker with gitlab-ce + postgres + redis)

storage2 (unencrypted, 2x6TB WD-Red)
	storage2
		public
			e-books
			media
				movies
				music
				pictures
				series
				tv-shows
				video
			software

backup1 (encrypted, external usb drive)
	backup1
		replicate
			<some dataset from zpool1>


Keep up the good work FreeNAS team!
 
Last edited:

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,194
Generally, the rule of thumb for when to choose a dataset instead of a directory is that the data is treated differently. Different snapshot schedule or retention policy, different owner, compression or no compression, (in the future) encryption or no encryption, ...
 

agartha79

Cadet
Joined
Aug 23, 2017
Messages
8
Generally, the rule of thumb for when to choose a dataset instead of a directory is that the data is treated differently. Different snapshot schedule or retention policy, different owner, compression or no compression, (in the future) encryption or no encryption, ...
Thanks for the quick reply! I've done my best effort to follow that rule of thumb with my layout presented here, but needless to say, it's not easy to fall upon one solution and stay with it. So before I complicate things more, I just need to know if there is any obvious mistakes with my layout that should be pointed out.

Here's a screenshot of my current setup:

ZQEne

It doesn't show anything here... so I include a link as well:
http://imgur.com/a/ZQEne

Notes:
- A symbolic link is created in user's home directory to access the main archive:
Code:
ln -s /mnt/storage1/archive /home/<username>/Archive
(Remember to include 'wide links = yes' in SMB Auxillary Parameters.

- The repo-gitlab-private is accessible from the VM using NFS.
 
Last edited:

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,194
The only obvious thing is that using subdatasets like you're suggesting may be unnecessary, but it won't hurt anything.
 

agartha79

Cadet
Joined
Aug 23, 2017
Messages
8
The only obvious thing is that using subdatasets like you're suggesting may be unnecessary, but it won't hurt anything.
That's good to hear. I'll stick to my current layout for now and see how things evolve. Have to move on :)
 
Status
Not open for further replies.
Top