]>
Commit | Line | Data |
---|---|---|
92f5a8d4 | 1 | 14.2.8 |
eafe8130 TL |
2 | ------ |
3 | ||
92f5a8d4 TL |
4 | * The following OSD memory config options related to bluestore cache autotuning can now |
5 | be configured during runtime: | |
eafe8130 | 6 | |
92f5a8d4 TL |
7 | - osd_memory_base (default: 768 MB) |
8 | - osd_memory_cache_min (default: 128 MB) | |
9 | - osd_memory_expected_fragmentation (default: 0.15) | |
10 | - osd_memory_target (default: 4 GB) | |
a8e16298 | 11 | |
92f5a8d4 | 12 | The above options can be set with:: |
81eedcae | 13 | |
92f5a8d4 | 14 | ceph config set global <option> <value> |
494da23a | 15 | |
92f5a8d4 TL |
16 | * The MGR now accepts 'profile rbd' and 'profile rbd-read-only' user caps. |
17 | These caps can be used to provide users access to MGR-based RBD functionality | |
18 | such as 'rbd perf image iostat' an 'rbd perf image iotop'. | |
eafe8130 | 19 | |
92f5a8d4 TL |
20 | * The configuration value ``osd_calc_pg_upmaps_max_stddev`` used for upmap |
21 | balancing has been removed. Instead use the mgr balancer config | |
22 | ``upmap_max_deviation`` which now is an integer number of PGs of deviation | |
23 | from the target PGs per OSD. This can be set with a command like | |
24 | ``ceph config set mgr mgr/balancer/upmap_max_deviation 2``. The default | |
25 | ``upmap_max_deviation`` is 1. There are situations where crush rules | |
26 | would not allow a pool to ever have completely balanced PGs. For example, if | |
27 | crush requires 1 replica on each of 3 racks, but there are fewer OSDs in 1 of | |
28 | the racks. In those cases, the configuration value can be increased. | |
eafe8130 | 29 | |
92f5a8d4 TL |
30 | * RGW: a mismatch between the bucket notification documentation and the actual |
31 | message format was fixed. This means that any endpoints receiving bucket | |
32 | notification, will now receive the same notifications inside an JSON array | |
33 | named 'Records'. Note that this does not affect pulling bucket notification | |
34 | from a subscription in a 'pubsub' zone, as these are already wrapped inside | |
35 | that array. | |
eafe8130 | 36 | |
92f5a8d4 TL |
37 | * Ceph will now issue a health warning if a RADOS pool as a ``pg_num`` |
38 | value that is not a power of two. This can be fixed by adjusting | |
39 | the pool to a nearby power of two:: | |
eafe8130 | 40 | |
92f5a8d4 | 41 | ceph osd pool set <pool-name> pg_num <new-pg-num> |
eafe8130 | 42 | |
92f5a8d4 | 43 | Alternatively, the warning can be silenced with:: |
eafe8130 | 44 | |
92f5a8d4 | 45 | ceph config set global mon_warn_on_pool_pg_num_not_power_of_two false |