[ceph-users] Data not distributed according to weights

2015-02-23 Thread Frank.Zirkelbach
Hello everyone, In my rather heterogeneous setup ... -1 54.36 root default -2 42.44 room 2.162 -4 6.09host le09091 3 2.03osd.3 up 1 1 2.03osd.1 up 1 9 2.03

Re: [ceph-users] Data not distributed according to weights

2015-02-23 Thread Christian Balzer
Hello, On Tue, 24 Feb 2015 06:28:25 + frank.zirkelb...@lew-verteilnetz.de wrote: > Hello everyone, > > In my rather heterogeneous setup ... > > -1 54.36 root default > -2 42.44 room 2.162 > -4 6.09host le09091 > 3 2.03

Re: [ceph-users] Data not distributed according to weights

2015-03-09 Thread Frank.Zirkelbach
Hello Christian, hello everyone, > That would be my guess, as a similar 4 node cluster here with a flat, host > only topology is having its data distributed as expected by the weights. To verify this, I eliminated the room buckets. Now all hosts are located within the root bucket. (Same tree as t

Re: [ceph-users] Data not distributed according to weights

2015-03-13 Thread Frank.Zirkelbach
Hello everyone, I think I figured out the reason, why in my setup the three small hosts are almost nearly full while there is plenty of free space on the only big one. (ceph osd tree output below) It's simply hitting the limit. The algorithm selects 3 hosts. Even if one copy was always on the bi