[Gluster-users] where is the free space?
Papp Tamas
tompos at martos.bme.hu
Wed Mar 6 11:53:27 UTC 2013
hi All,
Ubuntu 12.04 and glusterfs 3.3.1.
root at tipper:/data# df -h /data
Filesystem Size Used Avail Use% Mounted on
tipper:/data 2.0T 407G 1.6T 20% /data
root at tipper:/data# du -sh .
10G .
root at tipper:/data# du -sh /data
13G /data
It's quite confused.
I also tried to free up the space by stopping the machine (actually LXC VM) with no lock.
After umounting the space freed up right, though it took for a long time.
If I write directly to the fs, this behaviour is different.
This is a zfs the volume:
NAME PROPERTY VALUE SOURCE
tank/lxc/tipper/brick1 type filesystem -
tank/lxc/tipper/brick1 creation Fri Feb 15 14:26 2013 -
tank/lxc/tipper/brick1 used 16.4G -
tank/lxc/tipper/brick1 available 1.98T -
tank/lxc/tipper/brick1 referenced 16.4G -
tank/lxc/tipper/brick1 compressratio 1.11x -
tank/lxc/tipper/brick1 mounted yes -
tank/lxc/tipper/brick1 quota 2T local
tank/lxc/tipper/brick1 reservation none default
tank/lxc/tipper/brick1 recordsize 128K default
tank/lxc/tipper/brick1 mountpoint /tank/lxc/tipper/rootfs/mnt/brick1 local
tank/lxc/tipper/brick1 sharenfs off default
tank/lxc/tipper/brick1 checksum on default
tank/lxc/tipper/brick1 compression gzip local
tank/lxc/tipper/brick1 atime off inherited from
tank/lxc/tipper/brick1 devices on default
tank/lxc/tipper/brick1 exec on default
tank/lxc/tipper/brick1 setuid on default
tank/lxc/tipper/brick1 readonly off default
tank/lxc/tipper/brick1 zoned off default
tank/lxc/tipper/brick1 snapdir hidden default
tank/lxc/tipper/brick1 aclinherit restricted default
tank/lxc/tipper/brick1 canmount on default
tank/lxc/tipper/brick1 xattr on default
tank/lxc/tipper/brick1 copies 1 default
tank/lxc/tipper/brick1 version 5 -
tank/lxc/tipper/brick1 utf8only off -
tank/lxc/tipper/brick1 normalization none -
tank/lxc/tipper/brick1 casesensitivity sensitive -
tank/lxc/tipper/brick1 vscan off default
tank/lxc/tipper/brick1 nbmand off default
tank/lxc/tipper/brick1 sharesmb off default
tank/lxc/tipper/brick1 refquota none default
tank/lxc/tipper/brick1 refreservation none default
tank/lxc/tipper/brick1 primarycache all default
tank/lxc/tipper/brick1 secondarycache all default
tank/lxc/tipper/brick1 usedbysnapshots 0 -
tank/lxc/tipper/brick1 usedbydataset 16.4G -
tank/lxc/tipper/brick1 usedbychildren 0 -
tank/lxc/tipper/brick1 usedbyrefreservation 0 -
tank/lxc/tipper/brick1 logbias latency default
tank/lxc/tipper/brick1 dedup on local
tank/lxc/tipper/brick1 mlslabel none default
tank/lxc/tipper/brick1 sync standard default
tank/lxc/tipper/brick1 refcompressratio 1.11x -
tank/lxc/tipper/brick1 written 16.4G -
Do you have any idea, why that's happened?
10x
tamas
More information about the Gluster-users
mailing list