On Tue, Aug 13, 2013 at 01:16:05PM +0200, Valerio Pachera wrote: > 2013/8/13 Liu Yuan <namei.unix at gmail.com>: > > 1.5T, 1.3T, 1.4T looks kind of evenly distributed to me. > > What I concern about is not between hosts but between disks of node 3. > > > 0 465 GB 236 GB 229 GB 50% /mnt/sheep/dsk03 > > 1 1.8 TB 1.5 TB 314 GB 83% /mnt/sheep/dsk04 > > Yesterday was > > > 0 465 GB 236 GB 229 GB 48% /mnt/sheep/dsk03 > > 1 1.8 TB 1.5 TB 314 GB 79% /mnt/sheep/dsk04 > > /mnt/sheep/dsk04 is gettin full faster than the smaller /mnt/sheep/dsk03. > > > Can you check .stale in each disk is empty? > > They were all empty. > > > I think you'd better add a new node to the cluster. > > In theory I should be able to free space removing the snapshots :-) > -- This is only possible after we merge the Kazutaka's object reclaim patch set. Thanks Yuan |