- no_mon_update
- cluster_create
with_legacy: true
-- name: mon_delta_reset_interval
- type: float
- level: advanced
- desc: window duration for rate calculations in 'ceph status'
- fmt_desc: Seconds of inactivity before we reset the PG delta to 0. We keep
- track of the delta of the used space of each pool, so, for
- example, it would be easier for us to understand the progress of
- recovery or the performance of cache tier. But if there's no
- activity reported for a certain pool, we just reset the history of
- deltas of that pool.
- default: 10
- services:
- - mgr
- with_legacy: true
-- name: mon_stat_smooth_intervals
- type: uint
- level: advanced
- desc: number of PGMaps stats over which we calc the average read/write throughput
- of the whole cluster
- fmt_desc: Ceph will smooth statistics over the last ``N`` PG maps.
- default: 6
- services:
- - mgr
- min: 1
-- name: mon_pg_stuck_threshold
- type: int
- level: advanced
- desc: number of seconds after which pgs can be considered stuck inactive, unclean,
- etc
- long_desc: see doc/control.rst under dump_stuck for more info
- fmt_desc: Number of seconds after which PGs can be considered as
- being stuck.
- default: 1_min
- services:
- - mgr
-- name: mon_pg_warn_min_per_osd
- type: uint
- level: advanced
- desc: minimal number PGs per (in) osd before we warn the admin
- fmt_desc: Raise ``HEALTH_WARN`` if the average number
- of PGs per ``in`` OSD is under this number. A non-positive number
- disables this.
- default: 0
- services:
- - mgr
- name: mon_max_pg_per_osd
type: uint
level: advanced
default: 250
services:
- mgr
+ - mon
min: 1
-- name: mon_target_pg_per_osd
- type: uint
- level: advanced
- desc: Automated PG management creates this many PGs per OSD
- long_desc: When creating pools, the automated PG management logic will attempt to
- reach this target. In some circumstances, it may exceed this target, up to the
- ``mon_max_pg_per_osd`` limit. Conversely, a lower number of PGs per OSD may be
- created if the cluster is not yet fully utilised
- default: 100
- min: 1
-- name: mon_pg_warn_max_object_skew
- type: float
- level: advanced
- desc: max skew few average in objects per pg
- fmt_desc: Raise ``HEALTH_WARN`` if the average RADOS object count per PG
- of any pool is greater than ``mon_pg_warn_max_object_skew`` times
- the average RADOS object count per PG of all pools. Zero or a non-positive
- number disables this. Note that this option applies to ``ceph-mgr`` daemons.
- default: 10
- services:
- - mgr
-- name: mon_pg_warn_min_objects
- type: int
- level: advanced
- desc: 'do not warn below this object #'
- fmt_desc: Do not warn if the total number of RADOS objects in cluster is below
- this number
- default: 10000
- services:
- - mgr
-- name: mon_pg_warn_min_pool_objects
- type: int
- level: advanced
- desc: 'do not warn on pools below this object #'
- fmt_desc: Do not warn on pools whose RADOS object count is below this number
- default: 1000
- services:
- - mgr
-- name: mon_pg_check_down_all_threshold
- type: float
- level: advanced
- desc: threshold of down osds after which we check all pgs
- fmt_desc: Percentage threshold of ``down`` OSDs above which we check all PGs
- for stale ones.
- default: 0.5
- services:
- - mgr
- with_legacy: true
-- name: mon_cache_target_full_warn_ratio
- type: float
- level: advanced
- desc: issue CACHE_POOL_NEAR_FULL health warning when cache pool utilization exceeds
- this ratio of usable space
- fmt_desc: Position between pool's ``cache_target_full`` and ``target_max_object``
- where we start warning
- default: 0.66
- services:
- - mgr
- flags:
- - no_mon_update
- - cluster_create
- with_legacy: true
- name: mon_osd_full_ratio
type: float
level: advanced
- mon
see_also:
- ms_bind_msgr2
-- name: mon_warn_on_legacy_crush_tunables
- type: bool
- level: advanced
- desc: issue OLD_CRUSH_TUNABLES health warning if CRUSH tunables are older than mon_crush_min_required_version
- fmt_desc: Raise ``HEALTH_WARN`` when CRUSH tunables are too old (older than ``mon_min_crush_required_version``)
- default: true
- services:
- - mgr
- see_also:
- - mon_crush_min_required_version
- with_legacy: true
-- name: mon_crush_min_required_version
- type: str
- level: advanced
- desc: minimum ceph release to use for mon_warn_on_legacy_crush_tunables
- fmt_desc: The minimum tunable profile required by the cluster. See
- :ref:`CRUSH map tunables <crush-map-tunables>` for details.
- default: hammer
- services:
- - mgr
- see_also:
- - mon_warn_on_legacy_crush_tunables
- with_legacy: true
-- name: mon_warn_on_crush_straw_calc_version_zero
- type: bool
- level: advanced
- desc: issue OLD_CRUSH_STRAW_CALC_VERSION health warning if the CRUSH map's straw_calc_version
- is zero
- fmt_desc: Raise ``HEALTH_WARN`` when the CRUSH ``straw_calc_version`` is zero. See
- :ref:`CRUSH map tunables <crush-map-tunables>` for details.
- default: true
- services:
- - mgr
- with_legacy: true
-- name: mon_warn_on_osd_down_out_interval_zero
- type: bool
- level: advanced
- desc: issue OSD_NO_DOWN_OUT_INTERVAL health warning if mon_osd_down_out_interval
- is zero
- long_desc: Having mon_osd_down_out_interval set to 0 means that down OSDs are not
- marked out automatically and the cluster does not heal itself without administrator
- intervention.
- fmt_desc: Raise ``HEALTH_WARN`` when ``mon_osd_down_out_interval`` is zero. Having this
- option set to zero on the leader acts much like the ``noout`` flag. It's hard to figure
- out what's going wrong with clusters without the ``noout`` flag set but acting like that
- just the same, so we report a warning in this case.
- default: true
- services:
- - mgr
- see_also:
- - mon_osd_down_out_interval
- with_legacy: true
-- name: mon_warn_on_cache_pools_without_hit_sets
- type: bool
- level: advanced
- desc: issue CACHE_POOL_NO_HIT_SET health warning for cache pools that do not have
- hit sets configured
- fmt_desc: Raise ``HEALTH_WARN`` when a cache pool does not have the ``hit_set_type``
- value configured. See :ref:`hit_set_type <hit_set_type>` for more details.
- default: true
- services:
- - mgr
- with_legacy: true
-- name: mon_warn_on_pool_no_app
- type: bool
- level: dev
- desc: issue POOL_APP_NOT_ENABLED health warning if pool has not application enabled
- default: true
- services:
- - mgr
-- name: mon_warn_on_pool_pg_num_not_power_of_two
- type: bool
- level: dev
- desc: issue POOL_PG_NUM_NOT_POWER_OF_TWO warning if pool has a non-power-of-two
- pg_num value
- default: true
- services:
- - mon
-- name: mon_warn_on_pool_no_redundancy
- type: bool
- level: advanced
- desc: Issue a health warning if any pool is configured with no replicas
- fmt_desc: Raise ``HEALTH_WARN`` if any pool is configured with no replicas.
- default: true
- services:
- - mon
- see_also:
- - osd_pool_default_size
- - osd_pool_default_min_size
-- name: mon_allow_pool_size_one
- type: bool
- level: advanced
- desc: allow configuring pool with no replicas
- default: false
- services:
- - mon
-- name: mon_warn_on_misplaced
- type: bool
- level: advanced
- desc: Issue a health warning if there are misplaced objects
- default: false
- services:
- - mgr
- with_legacy: true
-- name: mon_warn_on_too_few_osds
- type: bool
- level: advanced
- desc: Issue a health warning if there are fewer OSDs than osd_pool_default_size
- default: true
- services:
- - mgr
- name: mon_warn_on_slow_ping_time
type: float
level: advanced
default: 0
services:
- mgr
+ - osd
see_also:
- mon_warn_on_slow_ping_ratio
- name: mon_warn_on_slow_ping_ratio
default: 0.05
services:
- mgr
+ - osd
see_also:
- osd_heartbeat_grace
- mon_warn_on_slow_ping_time
desc: Number of times to try sending a command directed at a specific monitor
default: 2
with_legacy: true
-- name: mon_max_pool_pg_num
- type: uint
- level: advanced
- default: 64_K
- fmt_desc: The maximum number of placement groups per pool.
-- name: mon_pool_quota_warn_threshold
- type: int
- level: advanced
- desc: percent of quota at which to issue warnings
- default: 0
- services:
- - mgr
-- name: mon_pool_quota_crit_threshold
- type: int
- level: advanced
- desc: percent of quota at which to issue errors
- default: 0
- services:
- - mgr
# whitespace-separated list of key=value pairs describing crush location
- name: crush_location
type: str
default: /usr/sbin/cephadm
services:
- mgr
+- name: mon_delta_reset_interval
+ type: float
+ level: advanced
+ desc: window duration for rate calculations in 'ceph status'
+ fmt_desc: Seconds of inactivity before we reset the PG delta to 0. We keep
+ track of the delta of the used space of each pool, so, for
+ example, it would be easier for us to understand the progress of
+ recovery or the performance of cache tier. But if there's no
+ activity reported for a certain pool, we just reset the history of
+ deltas of that pool.
+ default: 10
+ services:
+ - mgr
+ with_legacy: true
+- name: mon_stat_smooth_intervals
+ type: uint
+ level: advanced
+ desc: number of PGMaps stats over which we calc the average read/write throughput
+ of the whole cluster
+ fmt_desc: Ceph will smooth statistics over the last ``N`` PG maps.
+ default: 6
+ services:
+ - mgr
+ min: 1
+- name: mon_pool_quota_warn_threshold
+ type: int
+ level: advanced
+ desc: percent of quota at which to issue warnings
+ default: 0
+ services:
+ - mgr
+- name: mon_pool_quota_crit_threshold
+ type: int
+ level: advanced
+ desc: percent of quota at which to issue errors
+ default: 0
+ services:
+ - mgr
+- name: mon_cache_target_full_warn_ratio
+ type: float
+ level: advanced
+ desc: issue CACHE_POOL_NEAR_FULL health warning when cache pool utilization exceeds
+ this ratio of usable space
+ fmt_desc: Position between pool's ``cache_target_full`` and ``target_max_object``
+ where we start warning
+ default: 0.66
+ services:
+ - mgr
+ flags:
+ - no_mon_update
+ - cluster_create
+ with_legacy: true
+- name: mon_pg_check_down_all_threshold
+ type: float
+ level: advanced
+ desc: threshold of down osds after which we check all pgs
+ fmt_desc: Percentage threshold of ``down`` OSDs above which we check all PGs
+ for stale ones.
+ default: 0.5
+ services:
+ - mgr
+ with_legacy: true
+- name: mon_pg_stuck_threshold
+ type: int
+ level: advanced
+ desc: number of seconds after which pgs can be considered stuck inactive, unclean,
+ etc
+ long_desc: see doc/control.rst under dump_stuck for more info
+ fmt_desc: Number of seconds after which PGs can be considered as
+ being stuck.
+ default: 1_min
+ services:
+ - mgr
+- name: mon_pg_warn_min_per_osd
+ type: uint
+ level: advanced
+ desc: minimal number PGs per (in) osd before we warn the admin
+ fmt_desc: Raise ``HEALTH_WARN`` if the average number
+ of PGs per ``in`` OSD is under this number. A non-positive number
+ disables this.
+ default: 0
+ services:
+ - mgr
+- name: mon_pg_warn_max_object_skew
+ type: float
+ level: advanced
+ desc: max skew few average in objects per pg
+ fmt_desc: Raise ``HEALTH_WARN`` if the average RADOS object count per PG
+ of any pool is greater than ``mon_pg_warn_max_object_skew`` times
+ the average RADOS object count per PG of all pools. Zero or a non-positive
+ number disables this. Note that this option applies to ``ceph-mgr`` daemons.
+ default: 10
+ services:
+ - mgr
+- name: mon_pg_warn_min_objects
+ type: int
+ level: advanced
+ desc: 'do not warn below this object #'
+ fmt_desc: Do not warn if the total number of RADOS objects in cluster is below
+ this number
+ default: 10000
+ services:
+ - mgr
+- name: mon_pg_warn_min_pool_objects
+ type: int
+ level: advanced
+ desc: 'do not warn on pools below this object #'
+ fmt_desc: Do not warn on pools whose RADOS object count is below this number
+ default: 1000
+ services:
+ - mgr
+- name: mon_warn_on_misplaced
+ type: bool
+ level: advanced
+ desc: Issue a health warning if there are misplaced objects
+ default: false
+ services:
+ - mgr
+ with_legacy: true
+- name: mon_warn_on_pool_no_app
+ type: bool
+ level: dev
+ desc: issue POOL_APP_NOT_ENABLED health warning if pool has not application enabled
+ default: true
+ services:
+ - mgr
+- name: mon_warn_on_too_few_osds
+ type: bool
+ level: advanced
+ desc: Issue a health warning if there are fewer OSDs than osd_pool_default_size
+ default: true
+ services:
+ - mgr
+- name: mon_target_pg_per_osd
+ type: uint
+ level: advanced
+ desc: Automated PG management creates this many PGs per OSD
+ long_desc: When creating pools, the automated PG management logic will attempt to
+ reach this target. In some circumstances, it may exceed this target, up to the
+ ``mon_max_pg_per_osd`` limit. Conversely, a lower number of PGs per OSD may be
+ created if the cluster is not yet fully utilised
+ default: 100
+ min: 1
# min pgs per osd for reweight-by-pg command
- name: mon_reweight_min_pgs_per_osd
type: uint
default: 4_K
services:
- mon
+- name: mon_max_pool_pg_num
+ type: uint
+ level: advanced
+ default: 64_K
+ fmt_desc: The maximum number of placement groups per pool.
- name: mon_mgr_digest_period
type: int
level: dev
default: 0.0005
services:
- mon
+- name: mon_warn_on_cache_pools_without_hit_sets
+ type: bool
+ level: advanced
+ desc: issue CACHE_POOL_NO_HIT_SET health warning for cache pools that do not have
+ hit sets configured
+ fmt_desc: Raise ``HEALTH_WARN`` when a cache pool does not have the ``hit_set_type``
+ value configured. See :ref:`hit_set_type <hit_set_type>` for more details.
+ default: true
+ services:
+ - mon
+ with_legacy: true
+- name: mon_warn_on_pool_pg_num_not_power_of_two
+ type: bool
+ level: dev
+ desc: issue POOL_PG_NUM_NOT_POWER_OF_TWO warning if pool has a non-power-of-two
+ pg_num value
+ default: true
+ services:
+ - mon
+- name: mon_allow_pool_size_one
+ type: bool
+ level: advanced
+ desc: allow configuring pool with no replicas
+ default: false
+ services:
+ - mon
+- name: mon_warn_on_crush_straw_calc_version_zero
+ type: bool
+ level: advanced
+ desc: issue OLD_CRUSH_STRAW_CALC_VERSION health warning if the CRUSH map's straw_calc_version
+ is zero
+ fmt_desc: Raise ``HEALTH_WARN`` when the CRUSH ``straw_calc_version`` is zero. See
+ :ref:`CRUSH map tunables <crush-map-tunables>` for details.
+ default: true
+ services:
+ - mon
+ with_legacy: true
+- name: mon_warn_on_pool_no_redundancy
+ type: bool
+ level: advanced
+ desc: Issue a health warning if any pool is configured with no replicas
+ fmt_desc: Raise ``HEALTH_WARN`` if any pool is configured with no replicas.
+ default: true
+ services:
+ - mon
+ see_also:
+ - osd_pool_default_size
+ - osd_pool_default_min_size
+- name: mon_warn_on_osd_down_out_interval_zero
+ type: bool
+ level: advanced
+ desc: issue OSD_NO_DOWN_OUT_INTERVAL health warning if mon_osd_down_out_interval
+ is zero
+ long_desc: Having mon_osd_down_out_interval set to 0 means that down OSDs are not
+ marked out automatically and the cluster does not heal itself without administrator
+ intervention.
+ fmt_desc: Raise ``HEALTH_WARN`` when ``mon_osd_down_out_interval`` is zero. Having this
+ option set to zero on the leader acts much like the ``noout`` flag. It's hard to figure
+ out what's going wrong with clusters without the ``noout`` flag set but acting like that
+ just the same, so we report a warning in this case.
+ default: true
+ services:
+ - mon
+ see_also:
+ - mon_osd_down_out_interval
+ with_legacy: true
+- name: mon_warn_on_legacy_crush_tunables
+ type: bool
+ level: advanced
+ desc: issue OLD_CRUSH_TUNABLES health warning if CRUSH tunables are older than mon_crush_min_required_version
+ fmt_desc: Raise ``HEALTH_WARN`` when CRUSH tunables are too old (older than ``mon_min_crush_required_version``)
+ default: true
+ services:
+ - mon
+ see_also:
+ - mon_crush_min_required_version
+ with_legacy: true
+- name: mon_crush_min_required_version
+ type: str
+ level: advanced
+ desc: minimum ceph release to use for mon_warn_on_legacy_crush_tunables
+ fmt_desc: The minimum tunable profile required by the cluster. See
+ :ref:`CRUSH map tunables <crush-map-tunables>` for details.
+ default: hammer
+ services:
+ - mon
+ see_also:
+ - mon_warn_on_legacy_crush_tunables
+ with_legacy: true
- name: mon_warn_on_degraded_stretch_mode
type: bool
level: advanced
projects / ceph.git / commitdiff