.The least used OSDs are selected for increasing the weight
With this update, the least used OSD nodes are now selected for increasing the weight during the `reweight-by-utilization` process.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://rhn.redhat.com/errata/RHSA-2016-1972.html
[ubuntu@magna032 yum.repos.d]$ sudo ceph osd df ID WEIGHT REWEIGHT SIZE USE AVAIL %USE VAR 0 1.00000 1.00000 926G 198G 727G 21.45 1.13 1 0.89999 0.95554 926G 159G 766G 17.20 0.90 2 0.89999 1.00000 926G 167G 758G 18.09 0.95 3 0.89999 1.00000 926G 186G 739G 20.16 1.06 4 0.89999 1.00000 926G 191G 734G 20.72 1.09 5 0.89999 1.00000 926G 174G 751G 18.81 0.99 6 0.89999 1.00000 926G 202G 724G 21.82 1.15 7 0.89999 1.00000 926G 168G 758G 18.15 0.95 8 0.89989 0.79999 926G 137G 788G 14.90 0.78 TOTAL 8334G 1586G 6748G 19.03 MIN/MAX VAR: 0.78/1.15 STDDEV: 2.04 [ubuntu@magna032 yum.repos.d]$ sudo ceph osd test-reweight-by-utilization 101 0.05 no change moved 34 / 1925 (1.76623%) avg 213.889 stddev 23.8814 -> 17.3746 (expected baseline 13.7885) min osd.6 with 244 -> 237 pgs (1.14078 -> 1.10805 * mean) max osd.8 with 161 -> 176 pgs (0.752727 -> 0.822857 * mean) oload 101 max_change 0.05 max_change_osds 4 average 0.190327 overload 0.192230 osd.8 weight 0.799988 -> 0.849976 osd.6 weight 1.000000 -> 0.950012 osd.0 weight 1.000000 -> 0.950012 osd.1 weight 0.955536 -> 1.000000 Now based on the distance from the avg osds will be chosen i.e. higher the difference from avg utilization more chance of getting selected. Verified on 0.94.9-1.el7cp.x86_64