♣ mutex|compact|master|cache] print or configure basic namespace parameters
- ns stat [-a] [-m] [-n] [—reset] print namespace statistics
-a : break down by uid/gid
-m : display in monitoring format <key>=<value>
-n : display numerical uid/gid(s) —reset : reset namespace counters
- ns mutex [<option>] manage mutex monitoring.
Option can be:
—toggletime : toggle the timing
—toggleorder : toggle the order
—toggledeadlock : toggle deadlock check
—smplrate1 : set timing sample rate at 1% (default, no slow-down)
—smplrate10 : set timing sample rate at 10% (medium slow-down)
—smplrate100 : set timing sample rate at 100% (severe slow-down)
- ns compact off|on <delay> [<interval>] [<type>]
enable online compaction after <delay> seconds
<interval> : if >0 then compaction is repeated automatically
after so many seconds
<type> : can be ‘files’, ‘directories’ or ‘all’. By default only the file
changelog is compacted. The repair flag can be indicated by using
‘files-repair’, ‘directories-repair’ or ‘all-repair’.
- ns master [<option>] master/slave operations.
Option can be:
<master_hostname> : set hostname of MGM master RW daemon
—log : show master log
—log-clear : clean master log
—enable : enable the slave/master supervisor thread modifying stall/ redirectorion rules
—disable : disable supervisor thread
- ns recompute_tree_size <path>|cid:<decimal_id>|cxid:<hex_id> [—depth <val>] recompute the tree size of a directory and all its subdirectories
—depth : maximum depth for recomputation,
default 0 i.e no limit
- ns recompute_quotanode <path>|cid:<decimal_id>|cxid:<hex_id>
recompute the specified quotanode
- ns cache set|drop [-d|-f] [<max_num>] [<max_size>K|M|G…]
set the max number of entries or the max size of the cache. Use the ns stat command to see the current values.
set : update cache size for files or directories
drop : drop cached file and/or directory entries
-d : control the directory cache
-f : control the file cache
<max_num> : max number of entries
<max_size> : max size of the cache — not implemented yet
- ns cache drop-single-file <id of file to drop>
force refresh of the given FileMD by dropping it from the cache
- ns cache drop-single-container <id of container to drop>
force refresh of the given ContainerMD by dropping it from the cache
- ns max_drain_threads <num>
set the max number of threads in the drain pool, default 400, minimum 4
- ns reserve-ids <file id> <container id>
blacklist file and container IDs below the given threshold. The namespace will not allocate any file or container with IDs less than, or equal to the given blacklist thresholds.