.. highlight:: rst
.. index::
single: Filesystem API
fs
=====
fs ls
--------
List configured filesystems.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=ls
&eos.ruid=0
&eos.rgid=0
&mgm.format=json
[&mgm.outformat=l|m|e|d|io|fsck]
[&mgm.outhost=brief]
[&mgm.selection=<match>]
CLI syntax
++++++++++
.. code-block:: text
fs ls [-m|-l|-e|--io|--fsck|-d|--drain] [-s] [--brief|-b] [ [matchlist] ] :
list all filesystems in default output format. <space> is an optional substring match for the space name and can be a comma separated list
-m : list all filesystem parameters in monitoring format
--b,--brief : display host names without domain names
-l : display all filesystem parameters in long format
-e : display all filesystems in error state
--io : display all filesystems in IO output format
--fsck : display filesystem check statistics
-d,--drain : display all filesystems in drain or draindead status with drain progress and statistics
[matchlist] : [matchlist] can be just the name of the space to display or a comma seperated list of spaces e.g 'default,space'
[matchlist] : [matchlist] can be a grep style list to filter certain filesystems e.g. 'fs ls -d drain,bootfailure'
[matchlist] : [matchlist] can be a combination of space filter and grep e.g. 'fs ls -l space:default,drain,bootfailure'
fs add
--------
Add a new filesystem.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=add
&eos.ruid=0
&eos.rgid=0
[&mgm.fs.fsid=<fsid>]
&mgm.fs.uuid=<uuid>
&mgm.fs.node=<node-queue>
&mgm.fs.mountpoint=<mount-point>
&mgm.fs.space=<space>
&mgm.fs.configstatus=off|ro|rw
CLI syntax
++++++++++
.. code-block:: text
fs add [-m|--manual <fsid>] <uuid> <node-queue>|<host>[:<port>] <mountpoint> [<schedgroup>] [<status] :
add a filesystem and dynamically assign a filesystem id based on the unique identifier for the disk <uuid>
-m,--manual <fsid> : add with user specified <fsid> and <schedgroup> - no automatic assignment
<fsid> : numeric filesystem id 1..65535
<uuid> : arbitrary string unique to this particular filesystem
<node-queue> : internal EOS identifier for a node,port,mountpoint description ... /eos/<host>:<port>/fst e.g. /eos/myhost.cern.ch:1095/fst [you should prefer the host:port syntax]
<host> : fully qualified hostname where the filesystem is mounted
<port> : port where xrootd is running on the FST [normally 1095]
<mountpoint> : local path of the mounted filesystem e.g. /data
<schedgroup> : scheduling group where the filesystem should be inserted ... default is 'default'
<status> : file system status after the insert ... default is 'off', in most cases should be 'rw'
fs mv
--------
Move a filesystem between spaces/groups.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=mv
&eos.ruid=0
&eos.rgid=0
&mgm.space=<dst-space>
&mgm.fs.id=<fsid>|<space>
CLI syntax
++++++++++
.. code-block:: text
fs mv <src-fsid|src-space> <dst-schedgroup|dst-space> :
move a filesystem into a different scheduling group
<src-fsid> : source filesystem id
<src-space> : source space
<dst-schedgroup> : destination scheduling group
<dst-space> : destination space
If the source is a <space> a filesystem will be chosen to fit into the destination group or space.
If the target is a <space> : a scheduling group is auto-selected where the filesystem can be placed.
.. code-block:: text
fs config
---------
Configure filesystems.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=config
&eos.ruid=0
&eos.rgid=0
&mgm.fs.identifier=<fsid>
&mgm.fs.key=<key>
&mgm.fs.value=<value>
CLI syntax
++++++++++
.. code-block:: text
fs config <host>:<port><path>|<fsid>|<uuid> <key>=<value> :
configure filesystem parameter for a single filesystem identified by host:port/path, filesystem id or filesystem UUID.
fs config <fsid> configstatus=rw|wo|ro|drain|off :
<status> can be
rw : filesystem set in read write mode
wo : filesystem set in write-once mode
ro : filesystem set in read-only mode
drain : filesystem set in drain mode
off : filesystem set disabled
empty : filesystem is set to empty - possible only if there are no files stored anymorefs config <fsid> headroom=<size>
<size> can be (>0)[BMGT] : the headroom to keep per filesystem (e.g. you can write '1G' for 1 GB)
fs config <fsid> scaninterval=<seconds>:
configures a scanner thread on each FST to recheck the file & block checksums of all stored files every <seconds> seconds. 0 disables the scanning.
fs config <fsid> graceperiod=<seconds> :
grace period before a filesystem with an operation error get's automatically drained
fs config <fsid> drainperiod=<seconds> :
drain period a drain job is waiting to finish the drain procedure
fs rm
--------
Remove a configured filesystem.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=rm
&eos.ruid=0
&eos.rgid=0
&mgm.fs.id=<fsid|nodequeue>|&mgm.fs.node=<node>&mgm.fs.mountpoint=<mountpoint>
CLI syntax
++++++++++
.. code-block:: text
fs rm <fs-id>|<node-queue>|<mount-point>|<hostname> <mountpoint> :
remove filesystem configuration by various identifiers
fs boot
--------
Boot a filesystem.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=boot
&eos.ruid=0
&eos.rgid=0
&mgm.fsid=<fsid>|&mgm.node=<node>
[&mgm.fs.forcemgmsync=1]
CLI syntax
++++++++++
.. code-block:: text
fs boot <fs-id>|<node-queue>|* [--syncmgm]:
boot filesystem with ID <fs-id> or name <node-queue> or all (*)
--syncmgm : force an MGM resynchronization during the boot
fs dropdeletion
---------------
List configured filesystems.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=ls
&eos.ruid=0
&eos.rgid=0
&mgm.fs.id=<fsid>
CLI syntax
++++++++++
.. code-block:: text
fs dropdeletion <fs-id> :
allows to drop all pending deletions on <fs-id>
fs dumpmd
---------
Dump meta data of a filesystem.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=dumpmd
&eos.ruid=0
&eos.rgid=0
&mgm.fsid=<fsid
[&mgm.dumpmd.option=m]
[&mgm.dumpmd.fid=1]
[&mgm.dumpmd.size=1]
CLI syntax
++++++++++
.. code-block:: text
fs dumpmd [-s|-m] <fs-id> [-fid] [-path] :
dump all file meta data on this filesystem in query format
-s : don't printout keep an internal reference
-m : print the full meta data record in env format
-fid : dump only a list of file id's stored on this filesystem
-path : dump only a list of file names stored on this filesystem
fs satus
--------
Show status of a filesystem.
REST syntax
+++++++++++
.. code-block:: text
http://<host>:8000/proc/admin/ | root://<host>//proc/admin/
?mgm.cmd=fs
&mgm.subcmd=status
&eos.ruid=0
&eos.rgid=0
&mgm.fs.id=<fsid>|&mgm.fs.node=<nodeid>&mgm.fs.mountpoint=<mountpoint>
CLI syntax
++++++++++
.. code-block:: text
fs status [-l] <fs-id> :
returns all status variables of a filesystem and calculates the risk of data loss if this filesystem get's removed
fs status [-l] mount-point> :
as before but accepts the mount point as input parameters and set's host=<this host>
fs status [-l] <host> <mount-point> :
as before but accepts the mount point and hostname as input parameters
-l : list all files at risk and files which are offline