Zpool didn't grow

[Johhhn]

Hi, brief rundown:
Was: 6x3TB, now 6x5TB
Checked to make sure auto expand was on.
Did the offline/replace for each drive over the past week. Had zero problems, everything resilvered fine.
Today I did the online -e (then rebooted) just in case that would help.

Getting some funky size reports and hope you guys could help me.

Any thoughts on what's going on?

Code:

[root@huge] ~# zpool list
NAME   SIZE  ALLOC   FREE    CAP  DEDUP  HEALTH  ALTROOT
huge  27.2T  10.7T  16.5T    39%  1.00x  ONLINE  /mnt

[root@huge] ~# zfs list huge
NAME   USED  AVAIL  REFER  MOUNTPOINT
huge  7.15T  10.7T   384K  /mnt/huge

[root@huge] ~# df -h
Filesystem                                Size    Used   Avail Capacity  Mounted on
huge                                       10T    383k     10T     0%    /mnt/huge
huge/media                                 15T    5.1T     10T    32%    /mnt/huge/media
huge/tm                                    12T    2.0T     10T    16%    /mnt/huge/tm

 

[danb35]

How is your pool configured? RAIDZ2, Z3, etc.? The pool size itself looks right at 27 TiB; if you're using RAIDZ2 your usable space should be right about 18 TiB, which is what you're getting.

 

[Darren Myers]

using this ZFS calculator courtesy of @Bidule0hm it looks correct , as your previous storage should of been around 9.446TB , with the x6 5Tb hdds it should be 15.74TB which looks about right

 

[Johhhn]

How is your pool configured? RAIDZ2, Z3, etc.? The pool size itself looks right at 27 TiB; if you're using RAIDZ2 your usable space should be right about 18 TiB, which is what you're getting.

It's Z2, but what I don't understand is why it shows my pool huge under df -h to only show 10T for its size. And why my other two mount points show different sizes as well.

 

[Darren Myers]

Well it looks like you're using 7.1T , out of your ~17 making available 10...maybe someone else who knows a bit more can weigh in, but also remember theres overhead for ZFS and metadata

 

[danb35]

df doesn't provide valid information with ZFS, largely (or at least partially) because datasets are treated as separate filesystems. huge has 10 TiB free, and has 383K used, so its total size must be 10TiB. huge/media has the same 10 TiB free, but has 5.1 TiB used, so its total size must be 15 TiB. huge/tm has the same 10 TiB free, and 2 TiB used, so its total size must be 12 TiB. The 5.1 TiB in huge/media and the 2 TiB in huge/tm are part of huge, of course, but df is seeing those as three separate filesystems. Your pool did expand, and appears to be working as designed.

 

[Johhhn]

So the only way to know how much space i have total is to do some arithmetic? lol. seems odd that there isn't a total size or size that will show you how big your pool is.

So, in this case, based on the output below, I have 17.85T total, correct?

Code:

[root@huge] ~# zfs list huge
NAME   USED  AVAIL  REFER  MOUNTPOINT
huge  7.15T  10.7T   384K  /mnt/huge

df doesn't provide valid information with ZFS, largely (or at least partially) because datasets are treated as separate filesystems. huge has 10 TiB free, and has 383K used, so its total size must be 10TiB. huge/media has the same 10 TiB free, but has 5.1 TiB used, so its total size must be 15 TiB. huge/tm has the same 10 TiB free, and 2 TiB used, so its total size must be 12 TiB. The 5.1 TiB in huge/media and the 2 TiB in huge/tm are part of huge, of course, but df is seeing those as three separate filesystems. Your pool did expand, and appears to be working as designed.

 

[danb35]

The GUI will show you the total size.