zfs basics

[jdratlif]

I am very new to this, and I don't really understand what I'm doing. I was hoping to find a ZFS tutorial of some kind to explain what all these options are in NAS4Free.

I have 3 2 TiB HDDs. I want to put them into a raidz1 array. I have a 4th drive available as a backup, but I don't know if I should use it as a hot swap, so I've left it in the box for now.

I installed NAS4Free. I added the disks using Disks | Management | Import Disks.
I went to Disks | Format and formatted each disk as ZFS Storage Pool.
i went to Disks | ZFS | Pools | Virtual Device and added all 3 disks to a single RAIDZ1 array and named it HappyMonkey.
I went to Disks | ZFS | Pools | Management and created a single (i don't know what to call it -- mount point?), named it Giuseppe, and told it to use all the space (5.44 TB).

This seems to be working. I have been able to turn on CIFS and AFP and share data between Windows and Mac OS X Lion easily.

I know what a snapshot is, I think, though more information would be nice. What I don't know is what is a dataset or a volume. What would I use them for? Do I need them?

Is there a place I could read more about this? The wiki seems to have a lot of holes.

My system specs are:

AMD A4 2.7GHz Dual-Core
8 GB RAM
2 TiB HDD x3
4 GB Flash Drive (NAS4Free Embedded x64 9.1.0.1.344)

[torf74]

A dataset is the same as a filesystem if you were to consult the ZFS Administration Guide. A ZFS storage pool can contain a huge number of filesystems and this is useful for many reasons:

1. Space management and filesystem settings can be modified for each separate filesystem. If you have multiple users, you can allocate space and set other parameters accordingly and ZFS will manage it from within.
2. Snapshots are managed at the filesystem level. Snapshots can be rolled back on a filesystem level as well, and if you are ever in a position of needing to clone a snapshot or rolling back a snapshot, you will appreciate the ability to handle this in smaller chunks of disk space.

There are probably many other benefits, but these are the primary two that I can see in my everyday usage. I generally use a different dataset/filesystem for main folder organization. Music, Video, Documents, Backups, etc would all have their own filesystem under the main pool filesystem. I don't put anything directly under the main filesystem, since in the past I have experienced "holey filesystems" where some space is never released back into the pool. The only recourse at least at the time was to create a copy filesystem, copy over all data, and destroy the old holey filesystem, and then rename the new filesystem back to the old name.

One drawback to creating multiple filesystems is that that movement or copying between filesystems is fairly slow, as opposed to the nearly instantaneous movement and copying within a filesystem.

I use the following PDF file whenever I need to look up some information about ZFS:

http://www.google.com/url?sa=t&rct=j&q= ... kfu-_y7uKg

I don't know if it is the latest version, or even what versions are supported. On the old FreeNAS, there were plenty of times were the documentation was describing functionality not available in the ZFS pool version in FreeNAS .72. Since NAS4Free has been updated, it is probably more relavent. Maybe there is a newer PDF out there from Oracle, but I don't know how to find it. This one was the first hit in google.

[Onichan]

Just want to add that you should test your spare drive now rather then later. It would be bad to need the spare down the road and it ends up being a DOA.

[torf74]

Good point about verifying the status of your cold spares!

I did find a more up to date ZFS Admin Guide: http://www.google.com/url?sa=t&rct=j&q= ... 2wqYy24Ghg

Looks like it covers everything up to pool version 29. N4F is version 28 IIRC, so the vast majority of the guide should apply.