Why does an online change of the Ceph tunables "mon_osd_full_ratio" and "mon_osd_nearfull_ratio" does not take effect ?
Environment
- Red Hat Ceph Storage 1.2.3
- Red Hat Ceph Storage 1.2.3
- Ceph upstream firefly and Hammer
Issue
- Why does an online change of the Ceph tunables
mon_osd_full_ratioandmon_osd_nearfull_ratiodoes not take effect ?
Resolution
-
mon_osd_full_ratioandmon_osd_nearfull_ratioare not used for changing the values ofnearfull_ratioandfull_ratiofor PGs (OSDs). -
In place of
mon_osd_full_ratioandmon_osd_nearfull_ratio, please useceph pg set_full_ratioandceph pg set_nearfull_ratiocommands.
$ ceph pg set_full_ratio <value>
$ ceph pg set_nearfull_ratio <value>
Root Cause
-
Both
mon_osd_full_ratioandmon_osd_nearfull_ratioare used by Ceph to set the default values of filling ration and near-fill ratio, while creating placement groups. -
Instead of using these tunables, use
ceph pg set_full_ratio <value>andceph pg set_nearfull_ratio <value>.
Diagnostic Steps
$ ceph health detail
HEALTH_ERR 2 full osd(s); 140 near full osd(s)
osd.39 is full at 72%
osd.298 is full at 70%
$ ceph daemon /var/run/ceph/ceph-osd.39.asok config show | grep -i full_ratio
"mon_osd_full_ratio": "0.95",
"mon_osd_nearfull_ratio": "0.85",
"osd_backfill_full_ratio": "0.85",
"osd_pool_default_cache_target_full_ratio": "0.8",
"osd_failsafe_full_ratio": "0.97",
"osd_failsafe_nearfull_ratio": "0.9",
$ ceph daemon /var/run/ceph/ceph-osd.298.asok config show | grep -i full_ratio
"mon_osd_full_ratio": "0.95",
"mon_osd_nearfull_ratio": "0.85",
"osd_backfill_full_ratio": "0.85",
"osd_pool_default_cache_target_full_ratio": "0.8",
"osd_failsafe_full_ratio": "0.97",
"osd_failsafe_nearfull_ratio": "0.9",
This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.