Overview
ZFS stands for Zettabyte File system and to save you a lot of boring reading it is a file system which protects agains data corruption provide high redundancy and high availability. It was designed by Sun Microsystems and there are two implementations in nutshell:
- Oracle ZFS
- OpenProject ZFS
Both implementations are pretty close to each other and both are widely used in the IT world to provide availability FS with high protection and performance. ZFS is native to Oracle NAS Storage Appliance. With ZFS you can create a pool which you can share with out appliances through either NFS or Samba.
So let's create one pool:
Administration
We can create a pool, using the zfs command line as follows:
Manage a Pool
We all likes pools, especially with water and diving equipment. But let's create a ZPool now in RAID 2 with 10 disks and name it…..“tank” :)
- Create a pool
[root@hostname ~]# zpool create tank -O casesensitivity=mixed -O compression=lz4 raidz2 disk1 disk2 disk3 disk4 disk5 disk6 disk7 disk8 disk9 disk10 -f
The following command, will create you a pool comprised by 10 disks in raid-5 configuration with double parity. The pool name is called: “tank” and with compression lz4. You can choose one of these compression algorithms by the way :)
- On
- Off
- lzjb
- gzip
- gzip[1-9]
- zle
- lz4 ← The one used here and it is pretty good
You can also list the status of the created pool as follows:
- Pool status
[root@hostname ~]# zpool list -v NAME SIZE ALLOC FREE EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT tank 9.06T 1.94T 7.12T - 4% 21% 1.00x ONLINE - raidz2 9.06T 1.94T 7.12T - 4% 21% disk1 - - - - - - disk2 - - - - - - disk3 - - - - - - disk4 - - - - - - disk5 - - - - - - disk6 - - - - - - disk7 - - - - - - disk8 - - - - - - disk9 - - - - - - disk10 - - - - - -
If you are worried about the integrity of the pool, you can of course perform integrity check and verify the data corruption as follows:
- Check Integrity
[root@hostname ~]# zpool scrub tank [root@hostname ~]# zpool status pool: tank state: ONLINE scan: scrub repaired 0B in 0h56m with 0 errors on Mon Aug 12 12:40:46 2019 config: NAME STATE READ WRITE CKSUM tank ONLINE 0 0 0 raidz2-0 ONLINE 0 0 0 disk1 ONLINE 0 0 0 disk2 ONLINE 0 0 0 disk3 ONLINE 0 0 0 disk4 ONLINE 0 0 0 disk5 ONLINE 0 0 0 disk6 ONLINE 0 0 0 disk7 ONLINE 0 0 0 disk8 ONLINE 0 0 0 disk9 ONLINE 0 0 0 disk10 ONLINE 0 0 0 errors: No known data errors P.S. you can stop the scrubbing with: zpool scrub -s tank [root@hostname ~]# zpool scrub -s tank
After we have a pool, we can create a ZFS file system on it :)
Create file system
The ZFS will be shared either over SMB (Samba) or NFS(Network File system) to other machines. NFS is great way to setup Oracle RAC (with Direct NFS of course) and Samba is great way for a backup solution.
So let's create NFS :)
The command is fairly simple:
- Create ZFS File system
[root@hostname ~]# zfs create tank/backup
We can create also multiple aggregation of that file system, for example:
- Create Accumulative ZFS File system
[root@hostname ~]# zfs create tank/backup/servers/server_name
You can check the status of the file systems as well, very easy, as follows:
- List ZFS File Systems
[root@hostname ~]# zfs list NAME USED AVAIL REFER MOUNTPOINT tank 1.55T 5.45T 50.2K /tank tank/backup 1.55T 5.45T 52.6K /tank/backup tank/backup/zfsForServerOne 46.4G 5.45T 46.4G /tank/backup/forServerOne tank/backup/zfsForServerTwo 1.50T 5.45T 1.50T /tank/backup/forServerTwo
We can see that the total availability is 5.45 TBs, our of which are used: 1.55 TBs. 46.4 GBs on the file system for the first server and 1.5 TBs on the file system for the 2nd server.
We spoke, a lot about sharing and giving and stuff like that, but HOW we share of ZFS file system. So let's check it out :)
Share ZFS
In the past, sharing of the NFS was done via the following commands:
- Old way of sharing ZFS as NFS
[root@hostname ~]# share -F nfs /tank [root@hostname ~]# cat /etc/dfs/sharetab /tank - nfs rw
That is the old way, so how is the new way. Well, we have to set who server can do what :) For example:
- New way of Sharing ZFS as NFS
[root@hostname ~]# zfs set sharenfs="rw=@serverOne/16,ro=@serverTwo/28" tank/backup/zfsForServerOne
What that does is:
- Give READ-WRITE access of ZFS: tank/backup/zfsForServerOne to Server One's IP with netmask: 16
- Give READ-ONLY access of ZFS: tank/backup/zfsForServerOne to Server Two's IP with netmask: 28
It is uselessly to say that the NFS port (111) should be allowed both ways :)
For example: server_one (111) ↔ (111) Oracle_NAS_Server (111) ↔ (111) server_two
Of course, such configuration (with READ-ONLY) isn't suitable for RAC, if you don't want to hit your head in the wall long time. If you wish, to do so, please go ahead and try to install RAC on that :)
That command will update also the sharetab file of course:
- Check the sharing settings
[root@tbp-slm-zfs01 ~]# cat /etc/dfs/sharetab /tank/backup/zfsForServerOne - nfs rw=@serverOne/16,ro=@serverTwo/28
After you start the NFS server on the Oracle NAS, you can try to mount it on serverOne and serverTwo. I will not cover this as I have done in other section, but it will look like a normal NFS:
- NFS on serverOne
[root@serverOne]# df Filesystem 1K-blocks Used Available Use% Mounted on devtmpfs 3985372 0 3985372 0% /dev tmpfs 4000116 0 4000116 0% /dev/shm tmpfs 4000116 33480 3966636 1% /run tmpfs 4000116 0 4000116 0% /sys/fs/cgroup /dev/mapper/volGroup-root 38774276 2454892 36319384 7% / /dev/sda1 999320 181048 749460 20% /boot tmpfs 800020 0 800020 0% /run/user/283600015 /dev/mapper/vg_mongo-lv_mongo 721426424 277759688 409058372 41% /app/data 172.23.4.101:/tank/backup/zfsForServerOne 7463166976 1611866112 5851300864 22% /apphome/backup
Cheers and enjoy :)
Compression Options
For more info you can check here :) Compression Algorithms