Release Notes
# GooseFS Master HA deployment configurationgoosefs.zookeeper.enabled=truegoosefs.zookeeper.address=<zk_quorum_1>:<zk_client_port>,<zk_quorum_2>:<zk_client_port>,<zk_quorum_3>:<zk_client_port>goosefs.underfs.hdfs.configuration=${HADOOOP_HOME}/etc/hadoop/core-site.xml:${HADOOP_HOME}/hadoop/etc/hadoop/hdfs-site.xmlgoosefs.master.journal.type=UFSgoosefs.master.journal.folder=hdfs://HDFSXXXX/goosefs# Master metadata storage method, recommended Heap + RocksDB method, supports metadata at scale of hundreds of millionsgoosefs.master.metastore=ROCKSgoosefs.master.metastore.block=ROCKSgoosefs.master.metastore.block.locations=ROCKSFor GooseFS metadata storage directory, recommend choose a directory on high IOPS storage media.goosefs.master.metastore.dir=/data/goosefs/metastore#Metadata exchange method, RANDOM is selected by default; if there is obvious recent hot data access, consider setting it to LRU;# goosefs.master.metastore.cache.type=LRU# Disable orphan block verification at startup to lower leader election timegoosefs.master.startup.block.integrity.check.enabled=false# You can also disable periodically validating orphan blocks logic depending on the actual situation# goosefs.master.periodic.block.integrity.check.interval=-1# If not used the TTL feature, can also consider disabling periodic file expire checkgoosefs.master.ttl.checker.interval.ms=-1Can consider disabling replica check to reduce Master overheadgoosefs.master.replication.check.interval=-1# Worker configurationgoosefs.worker.tieredstore.levels=1goosefs.worker.tieredstore.level0.alias=SSDgoosefs.worker.tieredstore.level0.dirs.path=/data1/goosefsWorker,/data2/goosefsWorker# Set the following Quota value according to actual conditions# goosefs.worker.tieredstore.level0.dirs.quota=2000G,2000Ggoosefs.worker.block.heartbeat.interval.ms=10secgoosefs.worker.tieredstore.free.ahead.bytes=134217728goosefs.user.block.worker.client.pool.max=512Security certification and user simulation relatedgoosefs.security.authorization.permission.enabled=truegoosefs.security.authentication.type=SIMPLE# goosefs.security.login.username=hadoop# goosefs.master.security.impersonation.hadoop.users=*# goosefs.security.login.impersonation.username=_HDFS_USER_# Client configurationgoosefs.user.client.transparent_acceleration.scope=GFS_UFSgoosefs.user.client.transparent_acceleration.enabled=truegoosefs.user.file.readtype.default=CACHEgoosefs.user.file.writetype.default=CACHE_THROUGHgoosefs.user.metrics.collection.enabled=true
# GooseFS Master Raft deployment configurationgoosefs.master.rpc.addresses=<master1>:9200,<master2>:9200,<master3>:9200goosefs.master.embedded.journal.addresses=<master1>:9202,<master2>:9202,<master3>:9202Metadata checkpoint interval, defaults to 2000000, actual rate can be set based on metadata production speed in actual production environmentgoosefs.master.journal.checkpoint.period.entries=xxxx# GooseFS Journal data storage locationgoosefs.master.journal.folder=/data/goosefs/journal# Master metadata storage method, recommended Heap + RocksDB method, supports metadata at scale of hundreds of millionsgoosefs.master.metastore=ROCKSgoosefs.master.metastore.block=ROCKSgoosefs.master.metastore.block.locations=ROCKSFor GooseFS metadata storage directory, recommend choose a directory on high IOPS disks.goosefs.master.metastore.dir=/data/goosefs/metastore#Metadata exchange method, RANDOM is selected by default; if there is obvious recent hot data access, consider setting it to LRU;# goosefs.master.metastore.cache.type=LRU# Disable orphan block verification at startup to lower leader election timegoosefs.master.startup.block.integrity.check.enabled=false# You can also disable periodically validating orphan blocks logic depending on the actual situation# goosefs.master.periodic.block.integrity.check.interval=-1If not used, the TTL feature can be considered to disable periodic file expiration checkgoosefs.master.ttl.checker.interval.ms=-1# Can consider disabling replica check to reduce Master overheadgoosefs.master.replication.check.interval=-1# Worker configurationgoosefs.worker.tieredstore.levels=1goosefs.worker.tieredstore.level0.alias=SSDgoosefs.worker.tieredstore.level0.dirs.path=/data1/goosefsWorker,/data2/goosefsWorker# Set the following Quota value according to actual conditions# goosefs.worker.tieredstore.level0.dirs.quota=2000G,2000Ggoosefs.worker.block.heartbeat.interval.ms=10secgoosefs.worker.tieredstore.free.ahead.bytes=134217728goosefs.user.block.worker.client.pool.max=512# Security authentication and user simulation relatedgoosefs.security.authorization.permission.enabled=truegoosefs.security.authentication.type=SIMPLE# goosefs.security.login.username=hadoop# goosefs.master.security.impersonation.hadoop.users=*# goosefs.security.login.impersonation.username=_HDFS_USER_# Client configurationgoosefs.user.client.transparent_acceleration.scope=GFS_UFSgoosefs.user.client.transparent_acceleration.enabled=truegoosefs.user.file.readtype.default=CACHEgoosefs.user.file.writetype.default=CACHE_THROUGHgoosefs.user.metrics.collection.enabled=true
Esta página foi útil?
Você também pode entrar em contato com a Equipe de vendas ou Enviar um tíquete em caso de ajuda.
comentários