Free space question

[stratov]

Hi threre!

I'm new to FreeNAS, and setup a server for backups.

HW:
Intel Pentium G3420
16GB DDR3
6x3TB Seagate HDD
8GB USB Flash Drive for FreeNAS

SW:
FreeNAS-9.1.2.3-RELEASE-x64

So, it runs from USB flash drive, hdd drives configured at RAIDZ2, and i have 10.1T fo my data.
All i have used - is iSCSI mapped to Windows servers, i created 5 zvol, and assined each to separate iSCSI target
Total assigned space is 500G+500G+3T+1T+2T=7T

And few months later i've recognised, that the available free space is ~500G, and FreeNAS reported 90% capacity of the ZFS Volume

I checked volumes on servers, and none of them even half filled, but free space still leaking.

I'm afraid when there is no freespace FreeNAS just stop work or some another problem will occur

I tried scrub volume, reboot, SMART test with no luck

No snapshot used

Code:

[root@freenas ~]# zpool list                                             
NAME    SIZE  ALLOC  FREE    CAP  DEDUP  HEALTH  ALTROOT               
HMZFS  16.2T  14.8T  1.47T    90%  1.00x  ONLINE  /mnt 
           
[root@freenas ~]# zfs list                                               
NAME                  USED  AVAIL  REFER  MOUNTPOINT                   
HMZFS                10.2T  443G  320K  /mnt/HMZFS                   
HMZFS/.system        1.74M  443G  336K  /mnt/HMZFS/.system           
HMZFS/.system/cores    288K  443G  288K  /mnt/HMZFS/.system/cores     
HMZFS/.system/samba4  288K  443G  288K  /mnt/HMZFS/.system/samba4     
HMZFS/.system/syslog  871K  443G  871K  /mnt/HMZFS/.system/syslog     
HMZFS/Exchange_bckp    677G  443G  677G  -                             
HMZFS/distr            516G  830G  129G  -                             
HMZFS/headbackup      1.14T  443G  1.14T  -                             
HMZFS/mainbackup      5.59T  443G  5.59T  -                             
HMZFS/robocopy        2.32T  443G  2.32T  - 
 
[root@freenas ~]# zpool status                                                                                                      
  pool: HMZFS                                                                                                                       
 state: ONLINE                                                                                                                      
  scan: scrub repaired 0 in 16h16m with 0 errors on Wed Apr 23 04:49:52 2014                                                        
config:                                                                                                                             
                                                                                                                                    
        NAME                                            STATE     READ WRITE CKSUM                                                  
        HMZFS                                           ONLINE       0     0     0                                                  
          raidz2-0                                      ONLINE       0     0     0                                                  
            gptid/cf045937-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
            gptid/cfaf2093-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
            gptid/d03812cd-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
            gptid/d0b0dad0-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
            gptid/d12f9abc-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
            gptid/d1a63364-bbeb-11e3-b2e0-e03f49460087  ONLINE       0     0     0                                                  
                                                                                                                                    
errors: No known data errors   

 

[dlavigne]

Did you ever get this sorted?

 

[Yatti420]

Looks like the backups filled up.. Do you have snapshots running?

 

[stratov]

No, i dont use snapshots. Today Available space reached 0 and freenas bacame stuck. I deleted one of zvol (500Gb - Exchange_bckp), update it to 9.1.2.5. now Available space is 688Gb, but i think the behavior of system will be the same.

 

[cyberjock]

Well, zvols will take up more space than you actually allocate. The reason is that as you use it and the block size gets smaller you deal with more and more blocks storing that zvol which means even more checksums and such. This is why we don't recommend you create zvols or iscsi file extents that are more than 50% of your total pool because they will slowly fill up the pool.