The set command supports the following sub-commands:
This command will set the recovery profile to favor either the new client IO, recovery IO, or a balanced mode. The default is balanced
mode.
We can use the following built-in profile types:
- balanced
- high_client_ops
- high_recovery_ops
odf set recovery-profile high_client_ops
To verify the recovery profile setting run odf get recovery-profile.
This command allows updating the ceph osd full ratio in case ceph prevents the I/O operation on OSDs that reached the capacity specified. The default is 0.85.
Note: If the value is set too close to 1.0, the cluster will be unrecoverable if the OSDs are full and there is nowhere to grow.
odf set full 0.9
This command allows updating the ceph osd backfillfull ratio in case ceph will deny backfilling to the OSD that reached the capacity specified. The default value is 0.80.
Note: If the value is set too close to 1.0, the OSDs will be full and the cluster will not able to backfill.
odf set backfillfull 0.85
This command allows updating the ceph osd nearfull ratio in case Ceph returns the nearfull osds message when the cluster reaches the capacity specified. The default value is 0.75
Note: If the value is set too close to 1.0, the OSDs will be full and the cluster will not able to backfill.
odf set nearfull 0.8
The ceph
command helps update the Ceph configuration.
This command will set the log level for different ceph subsystems.
The debug_
prefix will be automatically added to the subsystem when enabling the logging.
odf set ceph log-level osd crush 20
Once the logging efforts are complete, restore the systems to their default or to a level suitable for normal operations.