OPTIONS¶
Configuration options:
DATA_PATH
where to store files with usage statistics and daemon
lock file
LABEL
the label of this chunkserver (for tiering)
WORKING_USER
user to run daemon as
WORKING_GROUP
group to run daemon as (optional - if empty then default
user group will be used)
SYSLOG_IDENT
name of process to place in syslog messages (default is
sfschunkserver)
LOCK_MEMORY
whether to perform mlockall() to avoid swapping out
sfschunkserver process (default is 0, i.e. no)
LIMIT_GLIBC_MALLOC_ARENAS
Linux only: limit glibc malloc arenas to given value -
prevents from using huge amount of virtual memory. This can influence
performance by reducing memory fragmentation and improving cache locality, but
it may also lead to contention and reduced parallelism in multi-threaded
applications. Use it in constrained memory environments, recommended values
are 4 or 8. (default is 0: disabled or let glibc decide)
MALLOC_TRIM_INTERVAL
interval in seconds at which the chunkserver will attempt
to release unused memory back to the OS using malloc_trim(). This can help
reduce the process’s memory footprint by returning free memory to the
system more frequently. Set to 0 to disable memory trimming entirely. (default
is 0, 0 = disabled)
NICE_LEVEL
nice level to run daemon with (default is -19 if
possible; note: process must be started as root to increase priority)
MASTER_HOST
address of SaunaFS master host to connect with (default
is sfsmaster)
MASTER_PORT
number of SaunaFS master port to connect with (default is
9420)
CLUSTER_ID
Only connect to metadata servers with this CLUSTER_ID.
Prevents accidental connections to wrong metadata servers. (default is
default)
MASTER_RECONNECTION_DELAY
delay in seconds before trying to reconnect to the master
server after disconnection (default is 5)
MASTER_TIMEOUT
timeout (in seconds) for the master server connection
(default is 60, minimum is 0.01)
MASTER_NR_OF_WORKERS
number of threads that the connection to master may use
to process operations on chunks, like create, duplicate, replicate and
truncate (default is 10, minimum is 2)
MASTER_REPLICATION_NR_OF_WORKERS
number of threads that the connection to master may use
to process replications operations on chunks. (default is 5, minimum in
1)
BIND_HOST
local address to use for connecting with the master
server (default is *, i.e. default local address)
CSSERV_LISTEN_HOST
IP address to listen on for client (mount) connections
(* means any)
CSSERV_LISTEN_PORT
port to listen on for client (mount) connections (default
is 9422)
HDD_CONF_FILENAME
alternative name of sfshdd.cfg file
HDD_LEAVE_SPACE_DEFAULT
free space threshold to set volume as 100% utisaued when
there is less than given amount of free space left (default is
"4GiB"). This number is always added to the used disk space reported
by chunkserver.
HDD_TEST_FREQ
chunk test period in seconds (default is 10)
HDD_CHECK_CRC_WHEN_READING
whether to check the CRC on every read operation (default
is 1)
HDD_CHECK_CRC_WHEN_WRITING
whether to check the CRC on every write operation. The
safest and recommended value is 1, disable it only to test maximum throughput.
(default is 1)
HDD_ADVISE_NO_CACHE
whether to remove each chunk from page when closing it to
reduce cache pressure generated by chunkserver (default is 0, i.e. no)
HDD_PUNCH_HOLES
if enabled then chunkserver detects zero values in chunk
data and frees corresponding file blocks (decreasing file system usage). This
option works only on Linux with file systems supporting punching holes (XFS,
ext4, Btrfs, tmpfs)
ENABLE_LOAD_FACTOR
if enabled, chunkserver will send periodical reports of
its I/O load to master, which will be taken into consideration when picking
chunkservers for I/O operations.
REPLICATION_BANDWIDTH_LIMIT_KBPS
limit how many kilobytes can be replicated from other
chunkservers to this chunkserver in every second (by default undefined, i.e.
no limits)
NR_OF_NETWORK_WORKERS
number of threads which handle (in a round-robin manner)
connections with clients (default is 4); these threads are responsible for
reading from sockets and coping data from internal buffers to sockets
NR_OF_HDD_WORKERS_PER_NETWORK_WORKER
number of threads that each network worker may use to do
disk operations like opening chunks, reading or writing them (default is
16)
BGJOBSCNT_PER_NETWORK_WORKER
maximum number of jobs that each network worker may use
for disk operations (default is 4000)
POLL_TIMEOUT_MS
Maximum amount of time in milliseconds that the polling
operation will wait for events. In the chunkservers, the same value is applied
for the events loop and for the network worker threads. Smaller values could
reduce latency at the cost of CPU usage (default: 50)
MAX_BLOCKS_PER_HDD_READ_JOB
maximum number of blocks read from the device in a single
read job. If possible, will try to make a single pread call (default is
8)
MAX_PARALLEL_HDD_READ_JOBS_PER_CS_ENTRY
maximum number of read jobs processed in parallel per
client connection (default is 16)
READ_AHEAD_KB
additional number of kilobytes which should be passed to
posix_fadvise(POSIX_FADV_WILLNEED) before reading data from a chunk (default
is 0, i.e. use posix_fadvise only with the amount of data that is really
needed; the value is aligned down to 64 KiB)
MAX_READ_BEHIND_KB
try to fix out-of-order read requests; the value tells
how much of skipped data to read if an offset of some read operation is
greater than the offset where the previous operation finished (default is 0,
i.e. don’t read any skipped data; the value is aligned down to 64
KiB)
PERFORM_FSYNC
call fsync() after a chunk is modified (default is 1,
i.e. enabled)
STAT_CHUNKS_AT_DISK_SCAN
verify that chunk metadata and data parts exists at scan
time. If set to 1, the data part is also checked to be multiple of
SFSBLOCKSIZE. Disabling this option accelerates the scan process, but does not
check the chunk files. Warning: use it with caution, only in scenarios where a
fast chunkserver restart is needed.(default is 1, i.e. enabled)
REPLICATION_TOTAL_TIMEOUT_MS
total timeout for single replication operation.
Replications that take longer than that are considered failed and are
immediately aborted (default: 60000)
REPLICATION_CONNECTION_TIMEOUT_MS
connection for single replication operation. If
connecting to another chunkserver takes longer than this timeout, the
operation is considered failed and is immediately aborted (default:
1000)
REPLICATION_WAVE_TIMEOUT_MS:: ADVANCED
timeout for single wave in replication. After this
timeout, next wave of read requests is sent to other chunkservers (default:
500)
CHUNK_TRASH_ENABLED (EXPERIMENTAL)
enables or disables the chunk trash feature. When
enabled, deleted chunks are moved to a trash directory instead of being
immediately removed. (Default: 0)
CHUNK_TRASH_EXPIRATION_SECONDS (EXPERIMENTAL)
specifies the timeout in seconds for chunks to remain in
the trash before being permanently deleted. (Default: 259200)
CHUNK_TRASH_FREE_SPACE_THRESHOLD_GB (EXPERIMENTAL)
sets the available space threshold in gigabytes. If the
available space on the disk falls below this threshold, the system will start
deleting older chunks from the trash to free up space. A suggested value is
about the 15% of the capacity of the smaller configured disk. (Default:
0)
CHUNK_TRASH_GC_BATCH_SIZE (EXPERIMENTAL)
defines the bulk size for the garbage collector when
processing chunks in the trash. This determines how many files are processed
in each garbage collection cycle. (Default: 1000)
CHUNK_TRASH_GC_SPACE_RECOVERY_BATCH_SIZE (EXPERIMENTAL)
[ADVANCED] The number of files to remove from the trash
in a single GC cycle, in case the disk is full and space needs to be
recovered. (Default: 100)
LOG_LEVEL
Setup logging. Uses the environment variable
SAUNAFS_LOG_LEVEL or config value LOG_LEVEL to determine logging level. Valid
log levels are
•trace
•debug
•info
•warn or warning
•err or error
•crit or critical
•off
