14.2.1 ------ * The default value for `mon_crush_min_required_version` has been changed from `firefly` to `hammer`, which means the cluster will issue a health warning if your CRUSH tunables are older than hammer. There is generally a small (but non-zero) amount of data that will move around by making the switch to hammer tunables; for more information, see :ref:`crush-map-tunables`. If possible, we recommend that you set the oldest allowed client to `hammer` or later. You can tell what the current oldest allowed client is with:: ceph osd dump | min_compat_client If the current value is older than hammer, you can tell whether it is safe to make this change by verifying that there are no clients older than hammer current connected to the cluster with:: ceph features The newer `straw2` CRUSH bucket type was introduced in hammer, and ensuring that all clients are hammer or newer allows new features only supported for `straw2` buckets to be used, including the `crush-compat` mode for the :ref:`balancer`. >=15.0.0 -------- * The RGW "num_rados_handles" has been removed. * If you were using a value of "num_rados_handles" greater than 1 multiply your current "objecter_inflight_ops" and "objecter_inflight_op_bytes" paramaeters by the old "num_rados_handles" to get the same throttle behavior. * Ceph now packages python bindings for python3.6 instead of python3.4, because EPEL7 recently switched from python3.4 to python3.6 as the native python3. see the `announcement _` for more details on the background of this change. * librbd now uses a write-around cache policy be default, replacing the previous write-back cache policy default. This cache policy allows librbd to immediately complete write IOs while they are still in-flight to the OSDs. Subsequent flush requests will ensure all in-flight write IOs are completed prior to completing. The librbd cache policy can be controlled via a new "rbd_cache_policy" configuration option. * librbd now includes a simple IO scheduler which attempts to batch together multiple IOs against the same backing RBD data block object. The librbd IO scheduler policy can be controlled via a new "rbd_io_scheduler" configuration option.