Is it possible to delete data from Ceph pools when the OSD disks becomes full, or hit the full_ratio of 95%?
Issue
-
Is it possible to delete data from Ceph pools when the OSD disks become full or hit the full_ratio of 95%?
-
This is specifically important for OpenStack implementations since there are no ways to delete volumes from OpenStack once the backend OSDs reach a full state.
-
Ceph OSD disks stop incoming I/O when it reaches 95% of its capacity. This will prevent any sort of I/O, even if it is to clean the disk off existing data to make space.
-
To work around this requires changing the
full_ratioof the OSD. -
It would be useful if there is a feature which can help in removing data from the pools in case the OSD disks hit full, at least for RBD.
Environment
-
Red Hat Ceph Storage 2.x
-
Red Hat OpenStack (using
Red Hat Ceph Storageas a backend for its instances)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
