Web-UI does not start + disk full


The NAS web-UI is not starting anymore.

If I log into the NAS, with ssh, I can check the journalctl logs to find warning No space left on device

ov 22 20:58:30 graph-nas kdumpctl[2315]: strip:/var/tmp/dracut.H4wdQM/initramfs/usr/lib/modules/4.10.6-1.el7.elrepo.x86_64/kernel/drivers/video/fbdev/core/strIkfEN[.note.gnu.build-id]: No space left on device

Checking the disk usage of the file system, I can find that the disk is full

btrfs filesystem df /
Data, single: total=52.25GiB, used=52.23GiB
System, single: total=8.00MiB, used=16.00KiB
Metadata, single: total=3.00GiB, used=2.60GiB
GlobalReserve, single: total=197.95MiB, used=0.00B

du -sh /
3.3G /

Could you help me to figure out how to recover the system ?

@dx_fred Welcome to the Rockstor community.

Yes, copy-on-write file systems can get pretty tricky when they are out of space. But you still have a little metadata left and the Global Reserve so you may be able to remove for example some older log files to get a little breathing room. Try the following to remove all but the last 2 days of system logs:

journalctl --vacuum-time=2d

Also note that the regular du command is not btrfs aware. You would need the btrfs specific du command thus:

btrfs fi du /

The output of this command before and after the journal clean up command above might be of use to others chipping in here later on.

And an overall look at your systems btrfs can be of use via:

btrfs fi show

As an aside, this is a super old kernel and version of Rockstor. You may want to consider re-installing with our latest available downloads from here: https://rockstor.com/dls.html

But it would be nice to first get you existing install in a state where you can do fresh backups and config saves etc before such a drastic move. But it would definitely be in you long term interests as we no longer release testing or stable updates for the now deprecated CentOS variant. Existing v3 pools and config can be imported as-is.

Hope that helps.

1 Like