/.system ate all my space!

[murple]

Hello,

Long time listener, first time caller.

I am running FreeNAS 9.2.1.3.

I have one ZFS pool which contains two disks which are mirrored.

A while ago I found out that my free space was basically 0. It seemed that automatic snapshots had eaten up all space. Disk was extremely slow and there were several thousands of snapshots.

I found a command which i ran it was:

Code:

zfs list -H -o name -t snapshot | xargs -n1 zfs destroy

This was successful and many gigabytes were freed.

Now I checked back again and see that free space is zero again. In the reporting graphs I can see that free space is decreasing, used space is not increasing.

I ran:

Code:

zfs list -o space

and result is:

Code:

NAME                 AVAIL   USED  USEDSNAP  USEDDS  USEDREFRESERV  USEDCHILD
pool                     0  1.78T         0   1.29T              0       506G
pool/.system             0   505G         0    168K              0       505G
pool/.system/cores       0   144K         0    144K              0          0
pool/.system/samba4      0  3.54M         0   3.54M              0          0
pool/.system/syslog      0   505G         0    505G              0          0
pool/jail                0  1.05G         0   1.05G              0          0
pool/plugins             0   303M         0    303M              0          0

USEDCHILD is hogging 505 GB of data in /.system. I have no idea what this is and couldn't find any information about it either when Googling.

I hope you can help me understand this.

Best regards,
murple

 

[Ericloewe]

Your syslog seems to be absolutely gigantic. What the hell happened here?

Did you happen to run very frequent snapshots on the system dataset?

 

[murple]

Yes probably had set up some stupid snapshot schedule and left it running for a year or two. Am I screwed now?

 

[anodos]

Hello,

Long time listener, first time caller.

I am running FreeNAS 9.2.1.3.

I have one ZFS pool which contains two disks which are mirrored.

A while ago I found out that my free space was basically 0. It seemed that automatic snapshots had eaten up all space. Disk was extremely slow and there were several thousands of snapshots.

I found a command which i ran it was:

Code:

zfs list -H -o name -t snapshot | xargs -n1 zfs destroy

This was successful and many gigabytes were freed.

Now I checked back again and see that free space is zero again. In the reporting graphs I can see that free space is decreasing, used space is not increasing.

I ran:

Code:

zfs list -o space

and result is:

Code:

NAME                 AVAIL   USED  USEDSNAP  USEDDS  USEDREFRESERV  USEDCHILD
pool                     0  1.78T         0   1.29T              0       506G
pool/.system             0   505G         0    168K              0       505G
pool/.system/cores       0   144K         0    144K              0          0
pool/.system/samba4      0  3.54M         0   3.54M              0          0
pool/.system/syslog      0   505G         0    505G              0          0
pool/jail                0  1.05G         0   1.05G              0          0
pool/plugins             0   303M         0    303M              0          0

USEDCHILD is hogging 505 GB of data in /.system. I have no idea what this is and couldn't find any information about it either when Googling.

I hope you can help me understand this.

Best regards,
murple

You can safely destroy the .system dataset "zfs destroy -fR pool/.system" and reboot. Fix your snapshot schedule.

 

[murple]

You can safely destroy the .system dataset "zfs destroy -fR pool/.system" and reboot. Fix your snapshot schedule.

Code:

NAME          AVAIL   USED  USEDSNAP  USEDDS  USEDREFRESERV  USEDCHILD
pool           505G  1.29T         0   1.29T              0      1.40G
pool/jail      505G  1.05G         0   1.05G              0          0
pool/plugins   505G   303M         0    303M              0          0

Worked wonders! Thank you very much!