- name: mds_alternate_name_max
type: size
level: advanced
- desc: set the maximum length of alternate names for dentries
+ desc: Set the maximum length of alternate names for dentries
default: 8_K
services:
- mds
- name: mds_fscrypt_last_block_max_size
type: size
level: advanced
- desc: maximum size of the last block without the header along with a truncate
+ desc: Maximum size of the last block without the header along with a truncate
request when the fscrypt is enabled.
default: 4_K
services:
- name: mds_numa_node
type: int
level: advanced
- desc: set mds's cpu affinity to a numa node (-1 for none)
+ desc: Set MDS CPU affinity to a NUMA node (-1 for none)
default: -1
services:
- mds
- name: mds_data
type: str
level: advanced
- desc: path to MDS data and keyring
+ desc: Path to MDS data and keyring
default: /var/lib/ceph/mds/$cluster-$id
services:
- mds
- name: mds_join_fs
type: str
level: basic
- desc: file system MDS prefers to join
- long_desc: This setting indicates which file system name the MDS should prefer to
+ desc: File system MDS prefers to join
+ long_desc: This setting indicates which CephFS file system the MDS should prefer to
join (affinity). The monitors will try to have the MDS cluster safely reach a
state where all MDS have strong affinity, even via failovers to a standby.
services:
- name: mds_cache_trim_interval
type: secs
level: advanced
- desc: interval in seconds between cache trimming
+ desc: Interval in seconds between cache trims
default: 1
services:
- mds
- name: mds_cache_quiesce_delay
type: millisecs
level: dev
- desc: delay before starting recursive quiesce inode operations
+ desc: Delay before starting recursive quiesce inode operations
default: 0
services:
- mds
- name: mds_cache_quiesce_splitauth
type: bool
level: advanced
- desc: allow recursive quiesce across auth boundaries
+ desc: Allow recursive quiesce across auth boundaries
default: true
services:
- mds
- name: mds_cache_release_free_interval
type: secs
level: dev
- desc: interval in seconds between heap releases
+ desc: Interval in seconds between heap releases
default: 10
services:
- mds
- name: mds_cache_memory_limit
type: size
level: basic
- desc: target maximum memory usage of MDS cache
+ desc: Target maximum memory usage of MDS cache
long_desc: This sets a target maximum memory usage of the MDS cache and is the primary
- tunable to limit the MDS memory usage. The MDS will try to stay under a reservation
+ tunable to limit MDS memory usage. The MDS will try to stay under a reservation
of this limit (by default 95%; 1 - mds_cache_reservation) by trimming unused metadata
in its cache and recalling cached items in the client caches. It is possible for
the MDS to exceed this limit due to slow recall from clients. The mds_health_cache_threshold
- name: mds_cache_reservation
type: float
level: advanced
- desc: amount of memory to reserve for future cached objects
- fmt_desc: The cache reservation (memory or inodes) for the MDS cache to maintain.
+ desc: Amount of memory to reserve for future cached objects
+ fmt_desc: The percentage of the configured cache size to reserve.
Once the MDS begins dipping into its reservation, it will recall
client state until its cache size shrinks to restore the
- reservation.
+ reservation. In other words, (1 - this value) is a cache fullness
+ high water mark above which client caps will be recalled.
default: 0.05
services:
- mds
- name: mds_health_cache_threshold
type: float
level: advanced
- desc: threshold for cache size to generate health warning
+ desc: Threshold for cache size to generate health warning
default: 1.5
services:
- mds
- name: mds_cache_mid
type: float
level: advanced
- desc: midpoint for MDS cache LRU
+ desc: Midpoint for MDS cache LRU
fmt_desc: The insertion point for new items in the cache LRU
(from the top).
default: 0.7
- name: mds_cache_trim_decay_rate
type: float
level: advanced
- desc: decay rate for trimming MDS cache throttle
+ desc: Decay rate for trimming MDS cache throttle
default: 1
services:
- mds
- name: mds_cache_trim_threshold
type: size
level: advanced
- desc: threshold for number of dentries that can be trimmed
+ desc: Threshold for number of dentries that can be trimmed
default: 256_K
services:
- mds
- name: mds_cache_quiesce_decay_rate
type: float
level: advanced
- desc: decay rate for quiescing inodes throttle
+ desc: Decay rate for quiescing inodes throttle
default: 1
services:
- mds
- name: mds_cache_quiesce_threshold
type: size
level: advanced
- desc: threshold for number of inodes that can be quiesced
+ desc: Threshold for number of inodes that can be quiesced
default: 512_K
services:
- mds
- name: mds_cache_quiesce_sleep
type: millisecs
level: advanced
- desc: sleep time for request after passing quiesce threshold
+ desc: Sleep time for requests after passing the quiesce threshold
default: 200
services:
- mds
- name: mds_max_file_recover
type: uint
level: advanced
- desc: maximum number of files to recover file sizes in parallel
+ desc: Maximum number of files for which to recover file sizes in parallel
default: 32
services:
- mds
- name: mds_dir_max_commit_size
type: int
level: advanced
- desc: maximum size in megabytes for a RADOS write to a directory
- fmt_desc: The maximum size of a directory update before Ceph breaks it into
+ desc: Maximum size in mebibytes for a RADOS write to a directory
+ fmt_desc: The maximum size of a directory update before the MDS splits it into
smaller transactions (MB).
default: 10
services:
- name: mds_dir_keys_per_op
type: int
level: advanced
- desc: number of directory entries to read in one RADOS operation
+ desc: Number of directory entries to read in one RADOS operation
default: 16384
services:
- mds
- name: mds_decay_halflife
type: float
level: advanced
- desc: rate of decay for temperature counters on each directory for balancing
+ desc: Rate of decay for temperature counters on each directory for balancing
default: 5
services:
- mds
- name: mds_beacon_interval
type: float
level: advanced
- desc: interval in seconds between MDS beacon messages sent to monitors
+ desc: Interval in seconds between MDS beacon messages sent to Monitors
default: 4
services:
- mds
- name: mds_beacon_grace
type: float
level: advanced
- desc: tolerance in seconds for missed MDS beacons to monitors
+ desc: Tolerance in seconds for missed MDS beacons to monitors
fmt_desc: The interval without beacons before Ceph declares an MDS laggy
- (and possibly replace it).
+ (and possibly replaces it).
default: 15
services:
- mds
- name: mds_heartbeat_reset_grace
type: uint
level: advanced
- desc: the basic unit of tolerance in how many circles in a loop, which will
- keep running by holding the mds_lock, it must trigger to reset heartbeat
+ desc: Tolerance in seconds for long-running MDS operations which do not
+ periodically reset the heartbeat timeout.
default: 1000
services:
- mds
- name: mds_heartbeat_grace
type: float
level: advanced
- desc: tolerance in seconds for MDS internal heartbeat
+ desc: Tolerance in seconds for MDS internal heartbeats
default: 15
services:
- mds
- name: mds_enforce_unique_name
type: bool
level: advanced
- desc: require MDS name is unique in the cluster
+ desc: Require unique MDS names
default: true
services:
- mds
- name: mds_session_blocklist_on_timeout
type: bool
level: advanced
- desc: blocklist clients whose sessions have become stale
+ desc: Blocklist clients whose sessions have become stale
default: true
services:
- mds
- name: mds_session_blocklist_on_evict
type: bool
level: advanced
- desc: blocklist clients that have been evicted
+ desc: Blocklist clients that have been evicted
default: true
services:
- mds
- name: mds_sessionmap_keys_per_op
type: uint
level: advanced
- desc: number of omap keys to read from the SessionMap in one operation
+ desc: Number of omap keys to read from the SessionMap in one operation
default: 1_K
services:
- mds
- name: mds_recall_max_caps
type: size
level: advanced
- desc: maximum number of caps to recall from client session in single recall
+ desc: Maximum number of caps to recall from a client session in single recall.
+ Note that this is an integer, though the default value may be displayed
+ with a B suffix.
default: 30000
services:
- mds
- name: mds_recall_max_decay_rate
type: float
level: advanced
- desc: decay rate for throttle on recalled caps on a session
+ desc: Decay rate for throttle on recalled caps on a session
default: 1.5
services:
- mds
- name: mds_recall_max_decay_threshold
type: size
level: advanced
- desc: decay threshold for throttle on recalled caps on a session
+ desc: Decay threshold for throttle on recalled caps on a session
default: 128_K
services:
- mds
- name: mds_recall_global_max_decay_threshold
type: size
level: advanced
- desc: decay threshold for throttle on recalled caps globally
+ desc: Decay threshold for throttle on recalled caps globally
default: 128_K
services:
- mds
- name: mds_recall_warning_threshold
type: size
level: advanced
- desc: decay threshold for warning on slow session cap recall
+ desc: Decay threshold for warning on slow session cap recall
default: 256_K
services:
- mds
- name: mds_recall_warning_decay_rate
type: float
level: advanced
- desc: decay rate for warning on slow session cap recall
+ desc: Decay rate for warning on slow session cap recall
default: 60
services:
- mds
- name: mds_session_cache_liveness_decay_rate
type: float
level: advanced
- desc: decay rate for session liveness leading to preemptive cap recall
+ desc: Decay rate for session liveness leading to preemptive cap recall
long_desc: This determines how long a session needs to be quiescent before the MDS
begins preemptively recalling capabilities. The default of 5 minutes will cause
10 halvings of the decay counter after 1 hour, or 1/1024. The default magnitude
- name: mds_session_cache_liveness_magnitude
type: size
level: advanced
- desc: decay magnitude for preemptively recalling caps on quiet client
+ desc: Decay magnitude for preemptively recalling caps on quiet client
long_desc: This is the order of magnitude difference (in base 2) of the internal
liveness decay counter and the number of capabilities the session holds. When
this difference occurs, the MDS treats the session as quiescent and begins recalling
- name: mds_session_cap_acquisition_decay_rate
type: float
level: advanced
- desc: decay rate for session readdir caps leading to readdir throttle
+ desc: Decay rate for session readdir caps leading to readdir throttle
long_desc: The half-life for the session cap acquisition counter of caps acquired
by readdir. This is used for throttling readdir requests from clients.
default: 30
- name: mds_session_cap_acquisition_throttle
type: uint
level: advanced
- desc: threshold at which the cap acquisition decay counter throttles
+ desc: Threshold at which the cap acquisition decay counter throttles
default: 100000
services:
- mds
- name: mds_session_max_caps_throttle_ratio
type: float
level: advanced
- desc: ratio of mds_max_caps_per_client that client must exceed before readdir may
+ desc: Ratio of mds_max_caps_per_client that client must exceed before readdir may
be throttled by cap acquisition throttle
default: 1.1
services:
- name: mds_cap_acquisition_throttle_retry_request_timeout
type: float
level: advanced
- desc: timeout in seconds after which a client request is retried due to cap acquisition
+ desc: Timeout in seconds after which a client request is retried due to cap acquisition
throttling
default: 0.5
services:
- name: mds_health_summarize_threshold
type: int
level: advanced
- desc: threshold of number of clients to summarize late client recall
+ desc: Threshold number of clients to summarize late client recall
default: 10
services:
- mds
- name: mds_reconnect_timeout
type: float
level: advanced
- desc: timeout in seconds to wait for clients to reconnect during MDS reconnect recovery
+ desc: Timeout in seconds to wait for clients to reconnect during MDS reconnect recovery
state
default: 45
services:
- name: mds_log_minor_segments_per_major_segment
type: uint
level: advanced
- desc: number of minor segments per major segment.
+ desc: Number of minor segments per major segment.
long_desc: The number of minor mds log segments since last major segment after which a major segment is started/logged.
default: 16
services:
- name: mds_file_blockdiff_max_concurrent_object_scans
type: uint
level: advanced
- desc: maximum number of concurrent object scans
+ desc: Maximum number of concurrent object scans
long_desc: Maximum number of concurrent listsnaps operations sent to RADOS.
default: 16
services:
projects / ceph.git / commitdiff