[SOLVED] Resize ZFS does not work

[Roi]

Hello all,

I tried to resize my ZFS array and followed this article:
http://harryd71.blogspot.de/2009/02/how ... world.html

Other hits on the Google search said similar things, so it seems to work or seemed to work.

In my case, I do not get a bigger array.

I changed my five 1.5TB drives to five 2TB drives and as described changed the drives one by one and resilvered between every change. This worked, the data is still there, but the size is also still the same.

Then I thought that encryption might be the problem and I did the same as with changing the drives. Killed the encryption on one drive, then resilvered, then the next and so on. So now the array is a non encrypted array.

Any ideas?

Cheers,
Roi

[raulfg3]

Any ideas?

yes, you need to turn ON Autoexpand.

see this post: http://forums.freebsd.org/showthread.php?t=20467

more info on oracle. http://docs.oracle.com/cd/E19253-01/819 ... index.html

[Roi]

Thank you for your reply. Sounds promising, but does not work.

basis:~# zpool status zfs-pool-1
pool: zfs-pool-1
state: ONLINE
scan: resilvered 230K in 0h0m with 0 errors on Mon May 6 08:09:49 2013
config:

NAME STATE READ WRITE CKSUM
zfs-pool-1 ONLINE 0 0 0
raidz1-0 ONLINE 0 0 0
ada2 ONLINE 0 0 0
ada3 ONLINE 0 0 0
ada4 ONLINE 0 0 0
ada5 ONLINE 0 0 0
ada6 ONLINE 0 0 0

errors: No known data errors
basis:~# zfs set autoexpand=on zfs-pool-1
cannot set property for 'zfs-pool-1': invalid property 'autoexpand'

I am using the latest (9.1.0.1 - Sandstorm (revision 636)) version of NAS4Free, so a too old ZFS cannot be the reason, right?

Cheers,
Roi

[raulfg3]

am using the latest (9.1.0.1 - Sandstorm (revision 636)) version of NAS4Free, so a too old ZFS cannot be the reason, right?

right, the reason is that you can't modify autoexpand.

try this:

Code: Select all

zpool set autoexpand=on zfs-pool-1

http://docs.oracle.com/cd/E19253-01/819 ... index.html

[Roi]

Thank you - but still, I do not have much luck... I also tried the export/import thing:

basis:~# zpool set autoexpand=on zfs-pool-1
basis:~# zpool list
NAME SIZE ALLOC FREE CAP DEDUP HEALTH ALTROOT
zfs-pool-1 6.81T 6.22T 607G 91% 1.00x ONLINE -
basis:~# zpool export zfs-pool-1
basis:~# zpool import zfs-pool-1
basis:~# zpool list
NAME SIZE ALLOC FREE CAP DEDUP HEALTH ALTROOT
zfs-pool-1 6.81T 6.22T 607G 91% 1.00x ONLINE -

[raulfg3]

first you need to check your pool (change NAS-A for zfs-pool-1 because NAS-A is my pool)

Code: Select all

rnas:~# zpool get all NAS-A
NAME   PROPERTY       VALUE       SOURCE
NAS-A  size           19T         -
NAS-A  capacity       68%         -
NAS-A  altroot        -           default
NAS-A  health         ONLINE      -
NAS-A  guid           17550733601870243541  default
NAS-A  version        28          default
NAS-A  bootfs         -           default
NAS-A  delegation     on          default
NAS-A  autoreplace    off         default
NAS-A  cachefile      -           default
NAS-A  failmode       wait        default
NAS-A  listsnapshots  off         default
NAS-A  autoexpand     off         default
NAS-A  dedupditto     0           default
NAS-A  dedupratio     1.00x       -
NAS-A  free           5.96T       -
NAS-A  allocated      13.0T       -
NAS-A  readonly       off         -
NAS-A  comment        -           default
NAS-A  expandsize     0           -
rnas:~#

as you can see I have autoexpand=off

now change to on and revise that is changed:

Code: Select all

rnas:~# zpool set autoexpand=on NAS-A
rnas:~# zpool get all NAS-A
NAME   PROPERTY       VALUE       SOURCE
NAS-A  size           19T         -
NAS-A  capacity       68%         -
NAS-A  altroot        -           default
NAS-A  health         ONLINE      -
NAS-A  guid           17550733601870243541  default
NAS-A  version        28          default
NAS-A  bootfs         -           default
NAS-A  delegation     on          default
NAS-A  autoreplace    off         default
NAS-A  cachefile      -           default
NAS-A  failmode       wait        default
NAS-A  listsnapshots  off         default
NAS-A  autoexpand     on          local
NAS-A  dedupditto     0           default
NAS-A  dedupratio     1.00x       -
NAS-A  free           5.96T       -
NAS-A  allocated      13.0T       -
NAS-A  readonly       off         -
NAS-A  comment        -           default
NAS-A  expandsize     0           -
rnas:~#

once done, try to do a scrub or export import to actualice size.

PD: revise that you change all disk, if only one disk is not changed by new ones, the entire pool can't grow

[Roi]

Thank you again for your reply.

I did change all disks from 1.5TB to 2TB, also I changed all encrypted disks to unencrypted disks (by reformating them one by one and relacing them into the pool).

This is what I get with the "zpool get all" command:

Code: Select all

basis:~# zpool get all zfs-pool-1
NAME        PROPERTY       VALUE       SOURCE
zfs-pool-1  size           6.81T       -
zfs-pool-1  capacity       91%         -
zfs-pool-1  altroot        -           default
zfs-pool-1  health         ONLINE      -
zfs-pool-1  guid           3820117156089091429  default
zfs-pool-1  version        28          default
zfs-pool-1  bootfs         -           default
zfs-pool-1  delegation     on          default
zfs-pool-1  autoreplace    off         default
zfs-pool-1  cachefile      -           default
zfs-pool-1  failmode       wait        default
zfs-pool-1  listsnapshots  off         default
zfs-pool-1  autoexpand     on          local
zfs-pool-1  dedupditto     0           default
zfs-pool-1  dedupratio     1.00x       -
zfs-pool-1  free           607G        -
zfs-pool-1  allocated      6.22T       -
zfs-pool-1  readonly       off         -
zfs-pool-1  comment        -           default
zfs-pool-1  expandsize     2.27T       -

So autoexpand is "on". I also find the last information interesting. Looks like the space I try to gain by replacing my disks.

As export/import did not work earlier, I started a scrub now. As you can imagine this will take a while.

[Roi]

The scrub is finished now, still no effect on the array... :-/ What is going on here? I am still curios what the "expandsize" parameter with the value 2.27T means and why it is not expanded to the new size.

[raulfg3]

sorry I do not understand why do not work, MUST WORK, sorry.

PD: Please if others users read this post , and have experience in growing up a zpool array, please try to help here.

[Roi]

GOT IT!!!

I found the solution here: http://zfsguru.com/forum/zfsgurusupport/482

Code: Select all

basis:~# zpool status zfs-pool-1
  pool: zfs-pool-1
 state: ONLINE
  scan: scrub repaired 0 in 8h53m with 0 errors on Thu May  9 05:00:04 2013
config:

        NAME        STATE     READ WRITE CKSUM
        zfs-pool-1  ONLINE       0     0     0
          raidz1-0  ONLINE       0     0     0
            ada2    ONLINE       0     0     0
            ada3    ONLINE       0     0     0
            ada4    ONLINE       0     0     0
            ada5    ONLINE       0     0     0
            ada6    ONLINE       0     0     0

errors: No known data errors
basis:~# zpool list
NAME         SIZE  ALLOC   FREE    CAP  DEDUP  HEALTH  ALTROOT
zfs-pool-1  6.81T  6.22T   607G    91%  1.00x  ONLINE  -
basis:~# zpool online -e zfs-pool-1 ada2 ada3 ada4 ada5 ada6
basis:~# zpool list
NAME         SIZE  ALLOC   FREE    CAP  DEDUP  HEALTH  ALTROOT
zfs-pool-1  9.06T  6.22T  2.84T    68%  1.00x  ONLINE  -

So the command zpool online -e finally did it for me.