List configured spaces.
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=ls
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
[&mgm.outformat=l|m|io|fsck]
[&mgm.selection=<match>]
space ls [-s] [-m|-l|--io|--fsck] [<space>] : list in all spaces or select only <space>. <space> is a substring match and can be a comma seperated list
-s : silent mode
-m : monitoring key=value output format
-l : long output - list also file systems after each space
--io : print IO satistics
--fsck : print filesystem check statistics
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=config
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
&mgm.space.name=<space>
&mgm.space.key=<key>
&mgm.space.value=<value>
space config <space-name> space.nominalsize=<value> : configure the nominal size for this space
space config <space-name> space.balancer=on|off : enable/disable the space balancer [default=off]
space config <space-name> space.balancer.threshold=<percent> : configure the used bytes deviation which triggers balancing [ default=20 (%) ]
space config <space-name> space.balancer.node.rate=<MB/s> : configure the nominal transfer bandwith per running transfer on a node [ default=25 (MB/s) ]
space config <space-name> space.balancer.node.ntx=<#> : configure the number of parallel balancing transfers per node [ default=2 (streams) ]
space config <space-name> space.converter=on|off : enable/disable the space converter [default=off]
space config <space-name> space.converter.ntx=<#> : configure the number of parallel conversions per space [ default=2 (streams) ]
space config <space-name> space.drainer.node.rate=<MB/s > : configure the nominal transfer bandwith per running transfer on a node [ default=25 (MB/s) ]
space config <space-name> space.drainer.node.ntx=<#> : configure the number of parallel draining transfers per node [ default=2 (streams) ]
space config <space-name> space.lru=on|off : enable/disable the LRU policy engine [default=off]
space config <space-name> space.lru.interval=<sec> : configure the default lru scan interval
space config <space-name> space.headroom=<size> : configure the default disk headroom if not defined on a filesystem (see fs for details)
space config <space-name> space.scaninterval=<sec> : configure the default scan interval if not defined on a filesystem (see fs for details)
space config <space-name> space.drainperiod=<sec> : configure the default drain period if not defined on a filesystem (see fs for details)
space config <space-name> space.graceperiod=<sec> : configure the default grace period if not defined on a filesystem (see fs for details)
space config <space-name> space.autorepair=on|off : enable auto-repair of faulty replica's/files (the converter has to be enabled too) => size can be given also like 10T, 20G, 2P ... without space before the unit
space config <space-name> space.geo.access.policy.write.exact=on|off : if 'on' use exact matching geo replica (if available) , 'off' uses weighting [ for write case ]
space config <space-name> space.geo.access.policy.read.exact=on|off : if 'on' use exact matching geo replica (if available) , 'off' uses weighting [ for read case ]
space config <space-name> fs.<key>=<value> : configure file system parameters for each filesystem in this space (see help of 'fs config' for details)
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=define
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
&mgm.space=<space>
&mgm.space.groumod=<groupmod>
&mgm.space.groupsize=<groupsize>
space define <space-name> [<groupsize> [<groupmod>]] : define how many filesystems can end up in one scheduling group <groupsize> [default=0]
=> <groupsize>=0 means, that no groups are built within a space, otherwise it should be the maximum number of nodes in a scheduling group
=> <groupmod> defines the maximun number of filesystems per node
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=reset
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
&mgm.space=<space>
&[mgm.option=egroup|mapping|drain|scheduledrain|schedulebalance]
space reset <space-name> [--egroup|mapping|drain|scheduledrain|schedulebalance]
: reset a space e.g. recompute the drain state machine
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=status
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
&mgm.space=<space>
space status <space-name> : print's all defined variables for space
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=space
&mgm.subcmd=set
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
&mgm.space=<space>
&mgm.space.state=on|off
space set <space-name> on|off : enables/disabels all groups under that space ( not the nodes !)