red-hat-storage/ocs-operator

Need ocs operator support OSDMaintenanceTimeout setting

par97 opened this issue · 1 comments

par97 commented

In a baremetal OCP cluster, we have ODF storage cluster using local disks, we wish to do storage node maintenance action, like node firmware upgrade, hardware replacement, etc.

In order to do that without any data rebuild, Rook has provided option of OSDMaintenanceTimeout setting inside DisruptionManagement struct of CephCluster spec.

But in OCS storagecluster CR, this DisruptionManagement/OSDMaintenanceTimeout option is not exposed, so set the OSDMaintenanceTimeout setting in storagecluster CR is not possible. Also if we update the cephcluster CR OSDMaintenanceTimeout setting, it would be changed back by ocs operator.

So I request to either expose this setting in StorageCluster CR spec, or make ocs operator do not change cephcluster CR OSDMaintenanceTimeout setting back to default, if we update this time out value.

Thanks.

par97 commented

Do not want to manually scale down ocs operator deployment, this is a known workaround. And ocs operator deployment seems could be auto scaled up after some time, not sure why.