DragonFly On-Line Manual Pages
mfsmaster.cfg(5) This is part of MooseFS mfsmaster.cfg(5)
NAME
mfsmaster.cfg - main configuration file for mfsmaster
DESCRIPTION
The file mfsmaster.cfg contains configuration of MooseFS master
process.
SYNTAX
Syntax is:
OPTION = VALUE
Lines starting with # character are ignored as comments.
OPTIONS
Configuration options:
WORKING_USER
user to run daemon as
WORKING_GROUP
group to run daemon as; optional value - if empty then default
user group will be used
SYSLOG_IDENT
name of process to place in syslog messages; default is
mfsmaster
LOCK_MEMORY
whether to perform mlockall() to avoid swapping out mfsmaster
process; default is 0, i.e. no
NICE_LEVEL
nice level to run daemon with; default is -19; note: process
must be started as root to increase priority, if setting of
priority fails, process retains the nice level it started with
FILE_UMASK
set default umask for group and others (user has always 0);
default is 027 - block write for group and block all for others
DATA_PATH
where to store metadata files and lock file
EXPORTS_FILENAME
alternate location/name of mfsexports.cfg file
TOPOLOGY_FILENAME
alternate location/name of mfstopology.cfg file
LICENCE_FILENAME
alternate location/name of mfslicence.bin file (pro version
only)
BACK_LOGS
number of metadata change log files (default is 50)
METADATA_DOWNLOAD_FREQ
how often (in hours) leader will download metadata from
followers (pro version only ; default is 24)
BACK_META_KEEP_PREVIOUS
number of previous metadata files to be kept (default is 1)
CHANGELOG_PRESERVE_SECONDS
how many seconds of change logs have to be preserved in memory
(default is 1800; this sets the minimum, actual number may be a
bit bigger due to logs being kept in 5k blocks; zero disables
extra logs storage)
MISSING_LOG_CAPACITY
how many missing chunks will be stored in master (up to
100*MISSING_LOG_CAPACITY bytes of memory will be allocated ;
default value is 100000)
MATOML_LISTEN_HOST
IP address to listen on for metalogger, masters and supervisors
connections (* means any)
MATOML_LISTEN_PORT
port to listen on for metalogger, masters and supervisors
connections
MASTER_RECONNECTION_DELAY
delay in seconds before next try to reconnect to master-leader
if not connected (pro version only ; default is 5)
MASTER_TIMEOUT
timeout in seconds for master-leader connections (pro version
only ; default is 10)
BIND_HOST
local address to use for connecting with master-leader (pro
version only ; default is *, i.e. default local address)
MATOCS_LISTEN_HOST
IP address to listen on for chunkserver connections (* means
any)
MATOCS_LISTEN_PORT
port to listen on for chunkserver connections
MATOCS_TIMEOUT
default timeout in seconds for master-chunkserver connection
(default is 10)
REPLICATIONS_DELAY_INIT
initial delay in seconds before starting replications (default
is 300)
CHUNKS_LOOP_MAX_CPS
Chunks loop shouldn't check more chunks per seconds than given
number (default is 100000)
CHUNKS_LOOP_MIN_TIME
Chunks loop shouldn't be done in less seconds than given number
(default is 300)
CHUNKS_SOFT_DEL_LIMIT
Soft maximum number of chunks to delete on one chunkserver
(default is 10)
CHUNKS_HARD_DEL_LIMIT
Hard maximum number of chunks to delete on one chunkserver
(default is 25)
CHUNKS_WRITE_REP_LIMIT
Maximum number of chunks to replicate to one chunkserver
(default is 2,1,1,4 - see NOTES)
CHUNKS_READ_REP_LIMIT
Maximum number of chunks to replicate from one chunkserver
(default is 10,5,2,5 - see NOTES)
CS_HEAVY_LOAD_THRESHOLD
Threshold for chunkserver load. (default is 100 - see NOTES)
CS_HEAVY_LOAD_RATIO_THRESHOLD
Threshold ratio for chunkserver load (default is 5.0 - see
NOTES)
CS_HEAVY_LOAD_GRACE_PERIOD
Defines how long chunkservers will remain in 'grace' mode
(default is 900 - see NOTES)
ACCEPTABLE_DIFFERENCE
Maximum difference between space usage of chunkservers
(deprecated, use ACCEPTABLE_PERCENTAGE_DIFFERENCE instead)
ACCEPTABLE_PERCENTAGE_DIFFERENCE
Maximum percentage difference between space usage of
chunkservers (default is 1 = 1%)
PRIORITY_QUEUES_LENGTH
Length of priority queues (for endangered, undergoal etc. chunks
- chunks that should be processed first - default is 1000000)
MATOCL_LISTEN_HOST
IP address to listen on for client (mount) connections (* means
any)
MATOCL_LISTEN_PORT
port to listen on for client (mount) connections
SESSION_SUSTAIN_TIME
How long to sustain a disconnected client session (in seconds;
default is 86400 = 1 day)
QUOTA_TIME_LIMIT
Time limit in seconds for soft quota (default is 604800 = 7
days)
NOTES
Chunks in master are tested in a loop. Speed (or frequency) is
regulated by two options CHUNKS_LOOP_MIN_TIME and CHUNKS_LOOP_MAX_CPS.
First defines minimal time between iterations of the loop and second
defines maximal number of chunk tests per second. Typically at the
beginning, when number of chunks is small, time is constant, regulated
by CHUNK_LOOP_MIN_TIME, but when number of chunks beccomes bigger then
time of loop can increase according to CHUNKS_LOOP_MAX_CPS.
Example: CHUNKS_LOOP_MIN_TIME is set to 300, CHUNKS_LOOP_MAX_CPS is set
to 100000 and there is 1000000 (one million) chunks in the system.
1000000/100000 = 10, which is less than 300, so one loop iteration will
take 300 seconds. With 1000000000 (one billion) chunks the system
needs 10000 seconds for one iteration of the loop.
Deletion limits are defined as 'soft' and 'hard' limit. When number of
chunks to delete increases from loop to loop, current limit can be
temporary increased above soft limit, but never above hard limit.
Replication limits are divided into four cases:
o first limit is for endangered chunks (chunks with only one copy)
o second limit is for undergoal chunks (chunks with number of copies
lower than specified goal)
o third limit is for rebalance between servers with space usage around
arithmetic mean
o fourth limit is for rebalance between other servers (very low or very
high space usage)
Usually first number should be grater than or equal to second, second
greater than or equal to third, and fourth greater than or equal to
third ( 1st >= 2nd >= 3rd <= 4th ). If one number is given, then all
limits are set to this number (for backward compatibility).
Whenever chunkserver load is higher than CS_HEAVY_LOAD_THRESHOLD and
CS_HEAVY_LOAD_RATIO_THRESHOLD times higher than average load, then
chunkserver is switched into 'grace' mode. Chunkserver stays in grace
mode for CS_HEAVY_LOAD_GRACE_PERIOD seconds.
COPYRIGHT
Copyright (C) 2016 Jakub Kruszona-Zawadzki, Core Technology Sp. z o.o.
This file is part of MooseFS.
MooseFS is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation, version 2 (only).
MooseFS is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License along
with MooseFS; if not, write to the Free Software Foundation, Inc., 59
Temple Place, Suite 330, Boston, MA 02111-1307 USA or visit
http://www.gnu.org/licenses/gpl-2.0.html
SEE ALSO
mfsmaster(8), mfsexports.cfg(5) mfstopology.cfg(5)
MooseFS 2.0.83-1 January 2016 mfsmaster.cfg(5)