ZFSLocalPoolCreation

Creating the Local ZFS Pool on nas local server for data storage

This is fun: we are about to unleash ZFS capabilities on our nas machine. Fasten your seatbelts and get ready.

Pool Creation on nas

• Suppose that NAS has /dev/sda as root, and that 3 disks will be used to create a raidz ZFS vdev, equivalent to a RAID-5 configuration.
• Install 3 unformatted disks. You do not even need to create a GPT partition table; ZFS tools will handle that too.

zpool create \    # command to create a pool 
-o ashift=12 \    # alignment of data, read somewhere (!)
-f \              # forces use of vdevs
zfspool \         # name of the pool
raidz \           # use raidz1 strategy, similar to RAID 5 of mdadm: you can loose 1 disk.
/dev/disk/by-id/ata-VBOX_HARDDISK_VB33e4cb68-e66806e3 \ serial # of /dev/sdb 
/dev/disk/by-id/ata-VBOX_HARDDISK_VB736b0a1e-d314dba7 \ serial # of /dev/sdc
/dev/disk/by-id/ata-VBOX_HARDDISK_VB3dfd818f-9b413e39 \ serial # of /dev/sdd

I created 3 (virtual) disks of 10GB each; the resulting pool will have 20GB of available space, since 10GB are used for parity.
Now we have a pool. Test it with zpool status:

user@nas:~$ zpool status
  pool: zfspool
 state: ONLINE
config:

        NAME                                       STATE     READ WRITE CKSUM
        zfspool                                    ONLINE       0     0     0
          raidz1-0                                 ONLINE       0     0     0
            ata-VBOX_HARDDISK_VB33e4cb68-e66806e3  ONLINE       0     0     0
            ata-VBOX_HARDDISK_VB736b0a1e-d314dba7  ONLINE       0     0     0
            ata-VBOX_HARDDISK_VB3dfd818f-9b413e39  ONLINE       0     0     0

errors: No known data errors

Setting pool property

• It is advisable to set the xattr=sa property at the pool level, so that it will be inherited from the filesystem(s) we will create on it. This prevents the storage of additional file attributes that would clutter differences and get in the way.

user@nas:~/src/zfs-backup$ sudo zfs set xattr=sa zfspool
user@nas:~/src/zfs-backup$ zfs get xattr
NAME                                   PROPERTY  VALUE  SOURCE
zfspool                                xattr     sa     local
zfspool/Documents                      xattr     sa     inherited from zfspool
zfspool/Documents@2024.07.26-00.48.47  xattr     sa     inherited from zfspool

Mounting the pool on /mnt/raid

By default the pool will be mounted on /zfspool :

user@nas:~$ zfs list
NAME      USED  AVAIL  REFER  MOUNTPOINT
zfspool   512K  19.0G   128K  /zfspool
user@nas:~$ df -h 
Filesystem                         Size  Used Avail Use% Mounted on
tmpfs                              392M  4.2M  388M   2% /run
/dev/mapper/ubuntu--vg-ubuntu--lv  6.1G  3.0G  2.8G  52% /
tmpfs                              2.0G     0  2.0G   0% /dev/shm
tmpfs                              5.0M     0  5.0M   0% /run/lock
/dev/sda2                          1.7G   95M  1.5G   6% /boot
tmpfs                              392M   12K  392M   1% /run/user/1000
zfspool                             20G  128K   20G   1% /zfspool

I like to mount it under /mnt/raid, so let's create that mountpoint.

user@nas:~$ sudo mkdir -p /mnt/raid
user@nas:~$ sudo zfs set mountpoint=/mnt/raid zfspool
user@nas:~$ zfs list
NAME      USED  AVAIL  REFER  MOUNTPOINT
zfspool   607K  19.0G   128K  /mnt/raid

Filesystem creation

Let's then create a filesystem that will hold our documents. I created many of them to separate music, photos, documents. Here we will just use one single filesystem; if you want more of them, just repeat for all the filesystems you wish to create.

user@nas:~$ sudo zfs create zfspool/Documents
user@nas:~$ zfs list
NAME                USED  AVAIL  REFER  MOUNTPOINT
zfspool             815K  19.0G   128K  /mnt/raid
zfspool/Documents   128K  19.0G   128K  /mnt/raid/Documents

Let's enable LZ4 compression on our filesystem. It is absolutely unnoticeable from the performance point of view and can save us some space with compressible files.

user@nas:~$ sudo zfs set compression=lz4 zfspool/Documents
user@nas:~$ zfs get compression zfspool/Documents
NAME               PROPERTY     VALUE           SOURCE
zfspool/Documents  compression  lz4             local

Now we have our filesystem mounted on /mnt/raid/Documents. We need to make it accessible to the user user:

user@nas:/mnt$ ls -la
total 9
drwxr-xr-x  3 root root 4096 Jul 20 23:32 .
drwxr-xr-x 23 root root 4096 Jul 20 23:37 ..
drwxr-xr-x  3 root root    3 Jul 20 23:37 raid
user@nas:/mnt$ sudo chown user:user /mnt/raid
user@nas:/mnt$ ls -la
total 9
drwxr-xr-x  3 root root 4096 Jul 20 23:32 .
drwxr-xr-x 23 root root 4096 Jul 20 23:37 ..
drwxr-xr-x  3 user user    3 Jul 20 23:37 raid
user@nas:/mnt$ cd raid
user@nas:/mnt/raid$ ls -la
total 5
drwxr-xr-x 3 user user    3 Jul 20 23:37 .
drwxr-xr-x 3 root root 4096 Jul 20 23:32 ..
drwxr-xr-x 2 root root    2 Jul 20 23:37 Documents
user@nas:/mnt/raid$ sudo chown user:user Documents
user@nas:/mnt/raid$ ls -la
total 5
drwxr-xr-x 3 user user    3 Jul 20 23:37 .
drwxr-xr-x 3 root root 4096 Jul 20 23:32 ..
drwxr-xr-x 2 user user    2 Jul 20 23:37 Documents
user@nas:/mnt/raid$ cd Documents
user@nas:/mnt/raid/Documents$ ls -la
total 1
drwxr-xr-x 2 user user 2 Jul 20 23:37 .
drwxr-xr-x 3 user user 3 Jul 20 23:37 ..
user@nas:/mnt/raid/Documents$ touch foo
user@nas:/mnt/raid/Documents$ ls -la
total 2
drwxr-xr-x 2 user user 3 Jul 20 23:46 .
drwxr-xr-x 3 user user 3 Jul 20 23:37 ..
-rw-rw-r-- 1 user user 0 Jul 20 23:46 foo
user@nas:/mnt/raid/Documents$ rm foo
user@nas:/mnt/raid/Documents$ ls -la
total 1
drwxr-xr-x 2 user user 2 Jul 20 23:46 .
drwxr-xr-x 3 user user 3 Jul 20 23:37 ..
user@nas:/mnt/raid/Documents$ 

Looks like we have a writeable Documents filesystem under our ZFS raidz vdev, with 20GB of available disk space.

Now head on to the backup server to create the backup pool.