ZFS
ZFS :- Zetta Byte filesystem is introduced on Solaris 10 Release.To
develop this filesystem cum volume manager,Sun Micro-systems had
spend lot of years and some billion dollars money. ZFS has many cool
features over traditional volume managers like SVM,VXVM.
MULTIPLE TERABYTE DISK SUPPORT IN ZFS, BUT SVM NOT SUPPORT.
ZFS 128BIT SUPPORT,BUT UFS 64BIT SUPPORT.
Advantages:-
1.Zpool Capacity of 256 zettabytes
2.ZFS snapshots,clones and Sending-receiving snapshots
3.Lightweight filesystem creation
4.Encryption
5.Software RAID
6.Data integrity
7.Integrated Volume management (No need an additional volume
manager)
Disadvantages :-
1.No way to reduce the zpool capacity
2. Re-silver takes more time in zpool raid.
(1) To Create a simple zpool :-
# zpool create <zpool_name> <device_name>
#zpool create szpool c1t2d0
# zpool list szpool
NAME SIZE ALLOC FREE CAP HEALTH ALTROOT
szpool 89M 97K 88.9M 0% ONLINE -
(2) To create a mirror zpool :-
# zpool create mzpool mirror c1t5d0 c1t6d0
# zpool list mzpool
NAME SIZE ALLOC FREE CAP HEALTH ALTROOT
mzpool 89M 97K 88.9M 0% ONLINE -
(3) To Create a raidz zpool :-
# zpool create rzpool raidz c1t2d0 c1t1d0 c1t8d0
# zpool list rzpool
NAME SIZE ALLOC FREE CAP HEALTH ALTROOT
rzpool 266M 176K 266M 0% ONLINE -
(4) In this second task, we are going to see how to create
new dataset under zpool.This is like creating new volumes
in vxvm :-
To create zfs dataset:You can see after creating the
volume,automatically dataset is mounted on /szpool/vol1
and zfs doesn’t require any vfstab entry for this :-
bash-3.00# zfs create szpool/vol1
bash-3.00# zfs list |grep szpool
szpool 105K 56.9M 21K /szpool
szpool/vol1 21K 56.9M 21K /szpool/vol1
(5) To set manual mount point:If you want to set specific
mount point for zfs dataset, use the below command :-
bash-3.00# zfs set mountpoint=/ora_vol1 szpool/vol1
bash-3.00# zfs list |grep szpool
szpool 115K 56.9M 22K /szpool
szpool/vol1 21K 56.9M 21K /ora_vol1
bash-3.00# df -h /ora_vol1
Filesystem size used avail capacity Mounted on
szpool/vol1 57M 21K 57M 1% /ora_vol1
(6) To share dataset(volume in vxvm,svm etc.) through
NFS:We can share the zfs dataset by modifying the zfs
attribute :-
bash-3.00# zfs get sharenfs szpool/vol1
NAME PROPERTY VALUE SOURCE
szpool/vol1 sharenfs off default
bash-3.00# zfs set sharenfs=on szpool/vol1
bash-3.00# zfs get sharenfs szpool/vol1
NAME PROPERTY VALUE SOURCE
szpool/vol1 sharenfs on local
(7) To compress dataset(volume in vxvm, svm etc.) ZFS
has default compression option.You can enable it using
zfs set command :-
bash-3.00# zfs get compression szpool/vol1
NAME PROPERTY VALUE SOURCE
szpool/vol1 compression off default
bash-3.00# zfs set compression=on szpool/vol1
bash-3.00# zfs get compression szpool/vol1
NAME PROPERTY VALUE SOURCE
szpool/vol1 compression on local
(8) To create dataset under dataset (sub volume in
volume) :-
bash-3.00# zfs create szpool/vol1/oraarch
bash-3.00# zfs list |grep ora
szpool/vol1 42K 56.9M 21K /ora_vol1
szpool/vol1/oraarch 21K 56.9M 21K /ora_vol1/oraarch
Diff btw Quota & Reservation :-
You can use the quota property to set a limit on the amount of
space a file system can use. In addition, you can use
the reservation property to guarantee that some amount of space
is available to a file system. Both properties apply to the dataset
they are set on and all descendents of that dataset.
(9) Setting reservation to dataset :-
bash-3.00# zfs set reservation=20M szpool/vol1/oraarch
bash-3.00# zfs get reservation szpool/vol1/oraarch
NAME PROPERTY VALUE SOURCE
szpool/vol1/oraarch reservation 20M local
bash-3.00# zfs list |grep ora
szpool/vol1 20.0M 36.9M 23K /ora_vol1
szpool/vol1/oraarch 21K 56.9M 21K /ora_vol1/oraarch
(10) By doing the above,you can see 20M is reserved for
oraarch and this space can’t be used by other
dataset.Setting quota to dataset :-
bash-3.00# zfs get quota szpool/vol1/oraarch
NAME PROPERTY VALUE SOURCE
szpool/vol1/oraarch quota none default
bash-3.00# zfs set quota=20M szpool/vol1/oraarch
bash-3.00# zfs get quota szpool/vol1/oraarch
NAME PROPERTY VALUE SOURCE
szpool/vol1/oraarch quota 20M local
bash-3.00# zfs list |grep ora
szpool/vol1 20.0M 36.9M 23K /ora_vol1
szpool/vol1/oraarch 21K 20.0M 21K /ora_vol1/oraarch
(11) To check the zpool status :-
bash-3.00# zpool status
pool: szpool
state: ONLINE
scrub: none requested
config:
NAME STATE READ WRITE CKSUM
szpool ONLINE 0 0 0
c1t3d0 ONLINE 0 0 0
errors: No known data errors
pool: rzpool
state: ONLINE
scrub: none requested
config:
NAME STATE READ WRITE CKSUM
rzpool ONLINE 0 0 0
raidz1-0 ONLINE 0 0 0
c1t2d0 ONLINE 0 0 0
c1t1d0 ONLINE 0 0 0
c1t8d0 ONLINE 0 0 0
errors: No known data errors
pool: mzpool
state: ONLINE
scrub: none requested
config:
NAME STATE READ WRITE CKSUM
mzpool ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
c1t5d0 ONLINE 0 0 0
c1t6d0 ONLINE 0 0 0
errors: No known data errors
(12) smcwebserver remote access :-
To access webbased zfs admin portal, use the following
link “https://system-name:6789/zfs”
In case if you are not getting the webpage in your server,
start the smcwebserver using the following command.
# /usr/sbin/smcwebserver start
If it is disbale,enable the service through the following command
# /usr/sbin/smcwebserver enable
sometimes smcwebserver will not able accessed remotly. In this case,
please follow the below steps to enable the remote access.
bash-3.00# svccfg -s svc:/system/webconsole setprop options/tcp_listen =
true
bash-3.00# svcadm refresh svc:/system/webconsole
bash-3.00# svcs -a |grep web
disabled Apr_01 svc:/application/management/webmin:default
online 1:01:02 svc:/system/webconsole:console
bash-3.00# /usr/sbin/smcwebserver restart
Restarting Oracle Java(TM) Web Console Version 3.1 ...
(13) If you want to create dataset with different mount
point, use the following command :-
# zpool create -m /export/zfs home c1t0d0
This is the example, szpool is an existing two-way mirror that is
transformed to a three-way mirror by attaching c2t1d0, the new device, to
the existing device, c1t1d0.
# zpool attach szpool c1t1d0 c2t1d0
# zpool detach zeepool c2t1d0
(14) To set auto-replace property on :-
# zpool set autoreplace=on wrkpool
To check property value
# zpool get autoreplace wrkpool
NAME PROPERTY VALUE SOURCE
wrkpool autoreplace on default
(15) Creating Emulated Volumes :-
# zfs create -V 5gb datapool/vol
To activate ZFS emulated volume as swap,
# swap -a /dev/zvol/dsk/datapool/vol
How to increase volume size:
# zfs set volsize=2g datapool/vol
(16) Creating ZFS Alternate Root Pools :-
# zpool create -R /mnt alt_rpool c0t0d0
Here we are giving ALT_ROOT pool name as alt_pool
# zfs list alt_pool
NAME USED AVAIL REFER MOUNTPOINT
morpheus 32.5K 33.5G 8K /mnt/alt_pool
Importing Alternate Root Pools:
# zpool import -R /mnt alt_pool
# zpool list alt_pool
NAME SIZE USED AVAIL CAP HEALTH ALTROOT
Morpheus 33.8G 68.0K 33.7G 0% ONLINE /mnt
# zfs list alt_pool
NAME USED AVAIL REFER MOUNTPOINT
Morpheus 32.5K 33.5G 8K /mnt/alt_pool
(17) To check the pool integity (Like fsck in UFS) :-
# zpool scrub datapool
i.e pool name is datapool
# zpool status -x
all pools are healthy
(18) To check the pool with detailed errors :-
# zpool status -v datapool
(19) Taking a Device Offline :-
# zpool offline datapool c0t0d0
bringing device ’c0t0d0’ offline
# zpool online datapool c0t0d0
bringing device ’c0t0d0’ online
(20) Replacing Devices :-
# zpool replace datapool c0t0d0 c0t0d1
In the above example, the previous device, c0t0d0, is replaced by c0t0d1
(21) IOSTAT :-
# zpool iostat
capacity operations bandwidth
pool used avail read write read write
———- —– —– —– —– —– —–
datapool 100G 20.0G 1.2M 102K 1.2M 3.45K
dozer 12.3G 67.7G 132K 15.2K 32.1K 1.20K
(22) Exporting a Pool :-
# zpool export datapool
cannot unmount ’/export/home/eschrock’: Device busy
# zpool export -f datapool
Determining Available Pools to Import
# zpool import
pool: datapool
id: 3824973938571987430916523081746329
Importing Pools
# zpool import datapool
To delete dataset
# zfs destroy datapool/home/tabriz
(23) To rename dataset :-
# zfs rename datapool/home/kustarz datapool/home/kustarz_old
First we need to mention existing volume name and new voume name
which you want to given.
To list all dataset under datapool/home/oracle3
zfs list -r datapool/home/oracle3
NAME USED AVAIL REFER MOUNTPOINT
datapool/home/oracle3 26.0K 4.81G 10.0K /datapool/home/oracle3
datapool/home/oracle3/projects 16K 4.81G 9.0K
/datapool/home/oracle3/projects
datapool/home/oracle3/projects/fs1 8K 4.81G 8K
/datapool/home/oracle3/projects/fs1
datapool/home/oracle3/projects/fs2 8K 4.81G 8K
/datapool/home/oracle3/projects/fs2
(24) Legacy Mount Points :-
# zfs set mountpoint=legacy datapool/home/eschrock
So that filesystem will mount automatically.We need to make entry in
vfstab to mount the FS
If you want to manually,use the following command
# mount -F zfs datapool/home/eschrock /mnt
The -a option can be used to mount all ZFS managed filesystems. Legacy
managed filesystems are not mounted.
# zfs mount -a
You can also share/unshare all ZFS filesystems on the system:
# zfs share -a
# zfs unshare datapool/home/tabriz
# zfs unshare -a
If the sharenfs property is off, then ZFS does not attempt to share or
unshare the filesystem at any time.
This allows the filesystem to be administered through traditional means
such as the /etc/dfs/dfstab file.
(25) Backing Up and Restoring ZFS Data :-
# zfs backup datapool/web1@111505 > /dev/rmt/0
# zfs restore datapool/test2@today < /dev/rmt/0
# zfs rename datapool/test datapool/test.old
# zfs rename datapool/test2 datapool/test
# zfs rollback datapool/web1@111505
cannot rollback to ’datapool/web1@111505’: more recent snapshots exist
use ’-r’ to force deletion of the following snapshots:
datapool/web1@now
# zfs rollback -r datapool/web1@111505
# zfs restore datapool/web1 < /dev/rmt/0
During the incremental restore process, the filesystem is unmounted and
cannot be accessed.
Remote Replication of a ZFS File System
# zfs backup datapool/sphere1@today | ssh newsys zfs restore
sandbox/restfs@today
restoring backup of datapool/sphere1@today
into sandbox/restfs@today …
restored 17.8Kb backup in 1 seconds (17.8Kb/sec)
# zfs send -I pool/fs@snap1 pool/clone@snapA > /snaps/fsclonesnap-I
(26) ZFS Snapshots and Clones :-
The following example creates a snapshot of datapool/home/ahrens that is
named friday.
# zfs snapshot datapool/home/ahrens@friday
# zfs destroy datapool/home/ahrens@friday
# zfs rename datapool/home/sphere1s@111205
datapool/home/sphere1s@today
Displaying and Accessing ZFS Snapshots
# ls /home/ahrens/.zfs/snapshot
tuesday wednesday thursday friday
Snapshots can be listed as follows:
# zfs list -t snapshot
NAME USED AVAIL REFER MOUNTPOINT
pool/home/ahrens@tuesday 13.3M – 2.13G –
zfs send -Rv wrkpool@0311 > /net/remote-system/rpool/snaps/wrkpool.0311
sending from @ to wrkpool@0311
sending from @ to wrkpool/swap@0311
sending from @ to wrkpool/dump@0311
sending from @ to wrkpool/ROOT@0311
sending from @ to wrkpool/ROOT/zfsnv109BE@zfsnv1092BE
sending from @zfsnv1092BE to wrkpool/ROOT/zfsnv109BE@0311
sending from @ to wrkpool/ROOT/zfsnv1092BE@0311
(27) ZFS Clones :-
# zfs clone pool/ws/gate@yesterday pool/home/ahrens/bug123
The following example creates a cloned work space from the
projects/newproject@today snapshot for a temporary user as
projects/teamA/tempuser and then sets properties on the cloned work
space.
# zfs snapshot projects/newproject@today
# zfs clone projects/newproject@today projects/teamA/tempuser
# zfs set sharenfs=on projects/teamA/tempuser
# zfs set quota=5G projects/teamA/tempuser
Destroying a Clone
ZFS clones are destroyed with the zfs destroy command.
# zfs destroy pool/home/ahrens/bug123
Clones must be destroyed before the parent snapshot can be destroyed
