Volume Size/Used/Available appears incorrect

[randomrat]

Hi All,

Complete FreeNAS new kid here and this is my first post! :) I've had a real interest in FreeNAS over the past few months and, after reading into it further, I have gone and built myself a dedicated storage system (please see sig for h/w details).

I've done the stage of creating a volume using the disk I currently have within the box, which are 4x 4TB WD Reds in a RAID-Z configuration. (I have read that implementing RAID-Z with disks of this size is highly un-recommended and that I should ideally use RAID-Z2 or Z3 - I understand the risk taken)

After setting up the RAID-Z volume, I had 10.4TB available to me which I was expecting (shown within the storage tab)...

The problem came shortly after when I began to copy data from my old Windows server to the datasets I had created within the pool...

I set the quota for Volume1 as 10.4T
I gave quotas for the following datasets:
Backup - 1.2T
Movies - 6.4T
Music - 400G
TV Shows - 2.4T
-------------------
Which all equals to 10.4T
-------------------

1) I then filled the datasets to a point with some data, and the space used within each dataset is not reflected in the volume's space used - it always reports at 0%. I can't seem to find anything in the manual that explains this.

2) Then, after copying more data across, all of a sudden the size of the volume dropped down from 10.4TB to 7.2TB...and then again to 6.7TB. I can still read/write/edit all the data within the volume.

3) Now the Used/Available/Size columns for all the datasets and the volume are showing values that I have no idea about or where they are coming from. I am using way more than 462GB and something more along the lines of 4TB.

I'm not sure whether this is an issue with FreeNAS or because I am doing something silly and just didn't read the manual thoroughly enough when creating volumes and datasets. I have read a few other threads with the same issue but nothing like this on 9.2.1 RELEASE.

Tonight I did an upgrade to 9.2.1.1-RELEASE Upgrade x64 seeing if it would resolve the issue but it didn't

After reading from other similar threads, doing "zfs list" will give much more accurate information. And reassuringly, if I add 4.13 and 6.24 together I get 10.37, which essentially equals 10.4TB.

Code:

[root@ODYSSEY ~]# zfs list                                                     
NAME                    USED  AVAIL  REFER  MOUNTPOINT                       
volume1                4.13T  6.24T  462G  /mnt/volume1                     
volume1/.samba4        2.25M  6.24T  2.25M  /mnt/volume1/.samba4             
volume1/.system        3.69M  6.24T  244K  /mnt/volume1/.system             
volume1/.system/cores    209K  6.24T  209K  /mnt/volume1/.system/cores       
volume1/.system/samba4  2.84M  6.24T  2.84M  /mnt/volume1/.system/samba4       
volume1/.system/syslog  413K  6.24T  413K  /mnt/volume1/.system/syslog       
volume1/Backup          209K  1.20T  209K  /mnt/volume1/Backup               
volume1/Movies          2.45T  3.95T  2.45T  /mnt/volume1/Movies               
volume1/Music          63.3G  337G  63.3G  /mnt/volume1/Music               
volume1/TV              1.17T  1.23T  1.17T  /mnt/volume1/TV  

I have also done "zpool list"

Code:

[root@ODYSSEY ~]# zpool list                                                   
NAME      SIZE  ALLOC  FREE    CAP  DEDUP  HEALTH  ALTROOT                   
volume1  14.5T  5.69T  8.81T    39%  1.00x  ONLINE  /mnt    

I don't really understand this, but does zpool list show the raw storage available?
So 16TB of disk is actually showing as 14.5TB?

And here is df -h...

Code:

[root@ODYSSEY ~]# df -h                                                       
Filesystem                Size    Used  Avail Capacity  Mounted on           
/dev/ufs/FreeNASs2a      926M    826M    26M    97%    /                     
devfs                    1.0k    1.0k      0B  100%    /dev                 
/dev/md0                  4.6M    3.3M    902k    79%    /etc                 
/dev/md1                  823k    2.0k    756k    0%    /mnt                 
/dev/md2                  149M    51M    86M    37%    /var                 
/dev/ufs/FreeNASs4        19M      3M    15M    16%    /data                 
volume1                  6.7T    462G    6.2T    7%    /mnt/volume1         
volume1/.samba4          6.2T    2.3M    6.2T    0%    /mnt/volume1/.samba4 
volume1/.system          6.2T    244k    6.2T    0%    /mnt/volume1/.system 
volume1/.system/samba4    6.2T    2.9M    6.2T    0%    /mnt/volume1/.system/sa
mba4                                                                           
volume1/.system/syslog    6.2T    418k    6.2T    0%    /mnt/volume1/.system/sy
slog                                                                           
volume1/.system/cores    6.2T    209k    6.2T    0%    /mnt/volume1/.system/co
res    

I hope I have been as concise as possible by giving all the relevant information.
Not too fussed if I have to delete the pool and destroy all data, but if I can avoid it I would obviously like to - I was warned that newbies can mess up the first few times so I still have everything on my old server.
I would be happy even if the answer was just some reassurance that I can still continue using my FreeNAS system without having to start all over again, and just wait for a bug fix if that's all it takes...
Really appreciate you reading this guys and gals!

 

[cyberjock]

Ok..

zfs list gives you actual user-available space.

zpool list gives you total raw disk space.. including parity. So its mostly useless for your intent.

df and du are broken with ZFS. They don't have an understanding of snapshots and other aspects of ZFS. So don't trust those numbers any more than you can throw them.

Now, the volume manager numbers do concern me. One of three things are going on:

1. Your browser is caching old data, so you aren't seeing current info.
2. You are using compression which is giving you amazing compression values.
3. There's a bug in FreeNAS' GUI.

Put a ticket in at bugs.freenas.org and provide everything you have here. Let the developers figure it out.

 

[cyberjock]

Aha.. here's that thread. Read this ticket...

https://bugs.freenas.org/issues/4419

 

[randomrat]

Thank you cyberjock.

Put a ticket in at bugs.freenas.org and provide everything you have here. Let the developers figure it out.

I have submitted a ticket and it appears to have been assigned to somebody to look at.
Before I went to bed I made a copy of the configuration file, did a fresh install of FreeNAS 9.2.1.1 RELEASE x64, (without the backed up config file) created my datasets, and began to copy some of my movies over again.

Aha.. here's that thread. Read this ticket...

Thanks for this. To clarify your point about "zfs list", it does represent the information more accurately and the numbers do add up, so im going to continue using FreeNAS as I was and shouldn't run into any issues from this. The number below wont match what's listed above as i'm still copying more movies over...

Code:

[root@ODYSSEY ~]# zfs list                                                   
NAME                  USED  AVAIL  REFER  MOUNTPOINT                         
tank                  679G  9.71T  267K  /mnt/tank                         
tank/.system        1.70M  9.71T  244K  /mnt/tank/.system                 
tank/.system/cores    209K  9.71T  209K  /mnt/tank/.system/cores           
tank/.system/samba4  872K  9.71T  872K  /mnt/tank/.system/samba4           
tank/.system/syslog  418K  9.71T  418K  /mnt/tank/.system/syslog           
tank/Backup          209K  2.00T  209K  /mnt/tank/Backup                   
tank/Games            209K  500G  209K  /mnt/tank/Games                   
tank/Media            679G  7.40T  244K  /mnt/tank/Media                   
tank/Media/Movies    679G  3.34T  679G  /mnt/tank/Media/Movies             
tank/Media/Music      209K  400G  209K  /mnt/tank/Media/Music             
tank/Media/TV        209K  3.00T  209K  /mnt/tank/Media/TV                 
tank/Programs        209K  500G  209K  /mnt/tank/Programs