Pool / Drive storage capacity not reporting properly

[originalprime]

Hello All,

Either I'm really tired and forgot all of my math skills or something isn't adding up correctly. I have three 1.5TB drives configured in a RAIDZ1 volume called "User_Data" which should equal something in the ballpark of 3TB usable space, formatting notwithstanding. Per the attached screenshot, it is only reporting 1.1TB of total space. It is also reporting 1.1TB of available space.

I only have one reservation set on a dataset called "Backup" of 500GB. The other datasets do not have reservations or quotas set. Something simply isn't adding up properly.

Is this a bug report, or can someone slap me back into reality and explain what's going on?

Thanks,
-Prime

 

[rabiat]

I've actually noticed the very same thing. I created a raidz1 volume consisting of 4x2TB disk, at first it correctly reported close to 6TB of usable space but all of a sudden it only reports 4TB...

zpool status -v
pool: volume1
state: ONLINE
scrub: none requested
config:

NAME STATE READ WRITE CKSUM
volume1 ONLINE 0 0 0
raidz1 ONLINE 0 0 0
gptid/262fbe20-dac2-11e0-8920-1c4bd65b9edc ONLINE 0 0 0
gptid/26dada8b-dac2-11e0-8920-1c4bd65b9edc ONLINE 0 0 0
gptid/277fbeef-dac2-11e0-8920-1c4bd65b9edc ONLINE 0 0 0
gptid/2818c354-dac2-11e0-8920-1c4bd65b9edc ONLINE 0 0 0

errors: No known data errors


-rabiat

 

[joeschmuck]

What version of FreeNAS are you running?

 

[originalprime]

Beg your pardon

@joeschmuck I'm sorry I left out that piece of info:

FreeNAS 8.01 stable.

 

[joeschmuck]

@originalprime
Stable, funny aren't you. I didn't see that issue on my machine. Did it happen right away or did it take time to drop? Also, if you reboot does the value return to 6TB or does it stick at 4TB?

Could you provide a "df -h" readout and/or possibly screen shots of the Reports tab?

 

[rabiat]

What version of FreeNAS are you running?

Sorry, I'm running 8.0.1-RC1. For me it took a few days or a week before it started to report 4TB instead of 6TB.

df -h
[...]
volume1 4.1T 2.0T 2.0T 50% /mnt/volume1


I can attach 'before and after' graph from the reporting page when I get home.

-rabiat

 

[joeschmuck]

What about the rest of what df returns? I'm actually interested in all of it. And before/after graphs would be great.

 

[rabiat]

What about the rest of what df returns? I'm actually interested in all of it. And before/after graphs would be great.

Here you go:

$ df -h

Filesystem Size Used Avail Capacity Mounted on
/dev/ufs/FreeNASs1a 927M 430M 422M 50% /
devfs 1.0K 1.0K 0B 100% /dev
/dev/md0 4.3M 3.6M 396K 90% /etc
/dev/md1 732K 16K 660K 2% /mnt
/dev/md2 75M 18M 51M 26% /var
/dev/ufs/FreeNASs4 20M 791K 17M 4% /data
volume1 4.1T 2.0T 2.0T 50% /mnt/volume1

 

[joeschmuck]

That is plain odd. I think you will be submitting a ticket against this problem. I might not be seeing the problem because my NAS doesn't run much but soon I'll be running continuously.

Have you run Scrub on it to see if something pops up? This take s along time so I'd recommend doing this before you go to bed.

The other thought is to save your configuration and reload FreeNAS 8.0.1-Release, then upload your configuration file, see if your 6TBs return.

 

[Milhouse]

Output from "zfs list"?

(Hint: Please put the otuput in {code} tags - makes interpreting it that much easier!)

 

[rabiat]

Output from "zfs list"?

(Hint: Please put the otuput in {code} tags - makes interpreting it that much easier!)

Code:

zfs list
NAME      USED  AVAIL  REFER  MOUNTPOINT
volume1  3.16T  2.02T  2.03T  /mnt/volume1

Code:

zpool list
NAME      SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
volume1  7.25T  4.35T  2.90T    60%  ONLINE  /mnt

 

[rabiat]

So..I have upgraded to the..erm.."stable" release now but the problem is still there.

I'm currently "scrubbing" the volume so we'll see tomorrow if something has changed, I doubt we will though.


-rabiat

 

[joeschmuck]

@Milhouse & rabiat

Since I'm learning too, the zfs list command results are basically saying there is 5.18TB of space on the hard drive for data under volume1? That doesn't tell me the unallocated space if any, correct?

EDIT: Okay, you just added the zpool list and now I'm confused. I'm learning as well.

 

[rabiat]

I'm very new to the world of ZFS, only played around with FreeNAS/ZFS for a few days so I don't know how to really interpret all the outputs :)

But I guess 2.02T would be the free/unallocated space.

 

[Milhouse]

@Milhouse & rabiat

Since I'm learning too, the zfs list command results are basically saying there is 5.18TB of space on the hard drive for data under volume1? That doesn't tell me the unallocated space if any, correct?

EDIT: Okay, you just added the zpool list and now I'm confused. I'm learning as well.

This is my take...

"zpool list" gives the raw device storage figures - this includes figures for any parity storage and associated metadata, and that's why the "zpool list" figures are so much higher than the "zfs list" results, as "zfs list" relates only to actual *usable* storage capacity (ie. the filesystem capacity).

So yes, it's a filesystem with 5.18TB of actual usable storage, built from a zpool consisting of disks that total 7.25T in capacity (ie. 4x"2TB" disks, although each disk is actually ~1,800MB).

 

[Milhouse]

But I guess 2.02T would be the free/unallocated space.

Yes, and to store 2.02TB of user data, ZFS will consume 2.90TB of disk space.

 

[originalprime]

I try to be a funny guy

@originalprime
Stable, funny aren't you. I didn't see that issue on my machine. Did it happen right away or did it take time to drop? Also, if you reboot does the value return to 6TB or does it stick at 4TB?

Could you provide a "df -h" readout and/or possibly screen shots of the Reports tab?

I do not know if the issue was immediate or if it dropped shortly after creation. Frankly I didn't catch it until I was loading data several hours after I had created the pool.

df -h readout:

Code:

Filesystem             Size    Used   Avail Capacity  Mounted on
/dev/ufs/FreeNASs1a    927M    429M    424M    50%    /
devfs                  1.0K    1.0K      0B   100%    /dev
/dev/md0               4.3M    3.6M    376K    91%    /etc
/dev/md1               732K     20K    656K     3%    /mnt
/dev/md2                75M     26M     43M    38%    /var
/dev/ufs/FreeNASs4      20M    1.4M     17M     8%    /data
User_Data              1.1T     39K    1.1T     0%    /mnt/User_Data
User_Data/Backup       500G    134G    366G    27%    /mnt/User_Data/Backup
User_Data/Documents    500G     99G    401G    20%    /mnt/User_Data/Documents
User_Data/Music        1.3T    194G    1.1T    14%    /mnt/User_Data/Music
User_Data/Pictures     1.3T    165G    1.1T    12%    /mnt/User_Data/Pictures
User_Data/Software     1.2T    106G    1.1T     8%    /mnt/User_Data/Software
User_Data/Videos       1.2T     89G    1.1T     7%    /mnt/User_Data/Videos
Media                  5.1T     32K    5.1T     0%    /mnt/Media
Media/Movies           5.3T    151G    5.1T     3%    /mnt/Media/Movies
Media/TV               5.2T     41G    5.1T     1%    /mnt/Media/TV

Weird, huh? All of my datasets excepting "Documents" and "Backup" report larger total size than their parent! Reboot has not solved the issue.

 

[Milhouse]

Weird, huh? All of my datasets excepting "Documents" and "Backup" report larger total size than their parent! Reboot has not solved the issue.

I don't suppose you enabled compression on the datasets other than Documents and Backup?

 

[originalprime]

zfs list & zpool list

@Milhouse

Here is my zfs list:

Code:

NAME                  USED  AVAIL  REFER  MOUNTPOINT
Media                 192G  5.14T  32.0K  /mnt/Media
Media/Movies          151G  5.14T   151G  /mnt/Media/Movies
Media/TV             41.3G  5.14T  41.3G  /mnt/Media/TV
User_Data            1.52T  1.15T  38.6K  /mnt/User_Data
User_Data/Backup      500G   366G   134G  /mnt/User_Data/Backup
User_Data/Documents   500G   401G  99.1G  /mnt/User_Data/Documents
User_Data/Music       194G  1.15T   194G  /mnt/User_Data/Music
User_Data/Pictures    165G  1.15T   165G  /mnt/User_Data/Pictures
User_Data/Software    106G  1.15T   106G  /mnt/User_Data/Software
User_Data/Videos     89.4G  1.15T  89.4G  /mnt/User_Data/Videos

Here is my zpool list:

Code:

NAME        SIZE   USED  AVAIL    CAP  HEALTH  ALTROOT
Media      8.12T   288G  7.84T     3%  ONLINE  /mnt
User_Data  4.06T  1.16T  2.91T    28%  ONLINE  /mnt

In the zpool, the pool status is closer to being correct. Per my original post, I am using three 1.5TB drives configured in RAIDZ1 for the "User_Data" pool, so 4.06TB is probably close if accounting for formatting... But my zfs list only shows 1.1TB. I am running a scrub now, but I will probably head on over to the bug reporting page anyway... Any other thoughts?

 

[originalprime]

@Milhouse

No compression enabled on any of them, at least not that I intentionally configured. The only non-default setting that I modified was to set a quota and reservation on the "Backup" dataset.