注册 登录  
 加关注
   显示下一条  |  关闭
温馨提示!由于新浪微博认证机制调整,您的新浪微博帐号绑定已过期,请重新绑定!立即重新绑定新浪微博》  |  关闭

mmicky 的博客

追逐刹那的惊艳

 
 
 

日志

 
 

Hbase缺省配置文件  

2013-11-13 11:26:01|  分类: Hbase |  标签: |举报 |字号 订阅

  下载LOFTER 我的照片书  |
hbase在0.95之后,分别有了hadoop1和hadoop2版了。在配置hbase的配置文件的时候,由于二进制的发布版所带的配置文件是空白的,给用户带来了配置的不便;事实上,在hbase的源码包中带有缺省的配置文件,如本人使用的hbase-0.96.0-src.tar.gz解压后在hbase-0.96.0/hbase-common/src/main/resources可以找到缺省的配置文件hbase-default.xml:

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<!--
/**
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance
 * with the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
-->

<!--
OVERVIEW

The important configs. are listed near the top.  You should change
at least the setting for hbase.tmp.dir.  Other settings will change
dependent on whether you are running hbase in standalone mode or
distributed.  See the hbase reference guide for requirements and
guidance making configuration.

This file does not contain all possible configurations.  The file would be
much larger if it carried everything. The absent configurations will only be
found through source code reading.  The idea is that such configurations are
exotic and only those who would go to the trouble of reading a particular
section in the code would be knowledgeable or invested enough in ever wanting
to alter such configurations, so we do not list them here.  Listing all
possible configurations would overwhelm and obscure the important.
-->

<configuration>
  <!--Configs you will likely change are listed here at the top of the file.
  -->
  <property >
    <name>hbase.tmp.dir</name>
    <value>${java.io.tmpdir}/hbase-${user.name}</value>
    <description>Temporary directory on the local filesystem.
    Change this setting to point to a location more permanent
    than '/tmp', the usual resolve for java.io.tmpdir, as the
    '/tmp' directory is cleared on machine restart.</description>
  </property>
  <property >
    <name>hbase.rootdir</name>
    <value>${hbase.tmp.dir}/hbase</value>
    <description>The directory shared by region servers and into
    which HBase persists.  The URL should be 'fully-qualified'
    to include the filesystem scheme.  For example, to specify the
    HDFS directory '/hbase' where the HDFS instance's namenode is
    running at namenode.example.org on port 9000, set this value to:
    hdfs://namenode.example.org:9000/hbase.  By default, we write
    to whatever ${hbase.tmp.dir} is set too -- usually /tmp --
    so change this configuration or else all data will be lost on
    machine restart.</description>
  </property>
  <property >
    <name>hbase.cluster.distributed</name>
    <value>false</value>
    <description>The mode the cluster will be in. Possible values are
      false for standalone mode and true for distributed mode.  If
      false, startup will run all HBase and ZooKeeper daemons together
      in the one JVM.</description>
  </property>
  <property>
    <name>hbase.zookeeper.quorum</name>
    <value>localhost</value>
    <description>Comma separated list of servers in the ZooKeeper ensemble
    (This config. should have been named hbase.zookeeper.ensemble).
    For example, "host1.mydomain.com,host2.mydomain.com,host3.mydomain.com".
    By default this is set to localhost for local and pseudo-distributed modes
    of operation. For a fully-distributed setup, this should be set to a full
    list of ZooKeeper ensemble servers. If HBASE_MANAGES_ZK is set in hbase-env.sh
    this is the list of servers which hbase will start/stop ZooKeeper on as
    part of cluster start/stop.  Client-side, we will take this list of
    ensemble members and put it together with the hbase.zookeeper.clientPort
    config. and pass it into zookeeper constructor as the connectString
    parameter.</description>
  </property>
  <!--The above are the important configurations for getting hbase up
    and running -->

  <property>
    <name>hbase.local.dir</name>
    <value>${hbase.tmp.dir}/local/</value>
    <description>Directory on the local filesystem to be used
    as a local storage.</description>
  </property>

  <!--Master configurations-->
  <property >
    <name>hbase.master.port</name>
    <value>60000</value>
    <description>The port the HBase Master should bind to.</description>
  </property>
  <property>
    <name>hbase.master.info.port</name>
    <value>60010</value>
    <description>The port for the HBase Master web UI.
    Set to -1 if you do not want a UI instance run.</description>
  </property>
  <property>
    <name>hbase.master.info.bindAddress</name>
    <value>0.0.0.0</value>
    <description>The bind address for the HBase Master web UI
    </description>
  </property>
  <property>
    <name>hbase.master.logcleaner.plugins</name>
    <value>org.apache.hadoop.hbase.master.cleaner.TimeToLiveLogCleaner</value>
    <description>A comma-separated list of LogCleanerDelegate invoked by
    the LogsCleaner service. These WAL/HLog cleaners are called in order,
    so put the HLog cleaner that prunes the most HLog files in front. To
    implement your own LogCleanerDelegate, just put it in HBase's classpath
    and add the fully qualified class name here. Always add the above
    default log cleaners in the list.</description>
  </property>
  <property>
    <name>hbase.master.logcleaner.ttl</name>
    <value>600000</value>
    <description>Maximum time a HLog can stay in the .oldlogdir directory,
    after which it will be cleaned by a Master thread.</description>
  </property>
  <property>
    <name>hbase.master.hfilecleaner.plugins</name>
    <value>org.apache.hadoop.hbase.master.cleaner.TimeToLiveHFileCleaner</value>
    <description>A comma-separated list of HFileCleanerDelegate invoked by
    the HFileCleaner service. These HFiles cleaners are called in order,
    so put the cleaner that prunes the most files in front. To
    implement your own HFileCleanerDelegate, just put it in HBase's classpath
    and add the fully qualified class name here. Always add the above
    default log cleaners in the list as they will be overwritten in
    hbase-site.xml.</description>
  </property>
  <property>
    <name>hbase.master.catalog.timeout</name>
    <value>600000</value>
    <description>Timeout value for the Catalog Janitor from the master to
    META.</description>
  </property>
  <property>
    <name>fail.fast.expired.active.master</name>
    <value>false</value>
    <description>If abort immediately for the expired master without trying
      to recover its zk session.</description>
  </property>
  <property>
    <name>hbase.master.dns.interface</name>
    <value>default</value>
    <description>The name of the Network Interface from which a master
      should report its IP address.</description>
  </property>
  <property>
    <name>hbase.master.dns.nameserver</name>
    <value>default</value>
    <description>The host name or IP address of the name server (DNS)
      which a master should use to determine the host name used
      for communication and display purposes.</description>
  </property>

  <!--RegionServer configurations-->
  <property>
    <name>hbase.regionserver.port</name>
    <value>60020</value>
    <description>The port the HBase RegionServer binds to.</description>
  </property>
  <property>
    <name>hbase.regionserver.info.port</name>
    <value>60030</value>
    <description>The port for the HBase RegionServer web UI
    Set to -1 if you do not want the RegionServer UI to run.</description>
  </property>
  <property>
    <name>hbase.regionserver.info.bindAddress</name>
    <value>0.0.0.0</value>
    <description>The address for the HBase RegionServer web UI</description>
  </property>
  <property>
    <name>hbase.regionserver.info.port.auto</name>
    <value>false</value>
    <description>Whether or not the Master or RegionServer
    UI should search for a port to bind to. Enables automatic port
    search if hbase.regionserver.info.port is already in use.
    Useful for testing, turned off by default.</description>
  </property>
  <property>
    <name>hbase.regionserver.handler.count</name>
    <value>30</value>
    <description>Count of RPC Listener instances spun up on RegionServers.
    Same property is used by the Master for count of master handlers.</description>
  </property>
  <property>
    <name>hbase.regionserver.msginterval</name>
    <value>3000</value>
    <description>Interval between messages from the RegionServer to Master
    in milliseconds.</description>
  </property>
  <property>
    <name>hbase.regionserver.optionallogflushinterval</name>
    <value>1000</value>
    <description>Sync the HLog to the HDFS after this interval if it has not
    accumulated enough entries to trigger a sync. Units: milliseconds.</description>
  </property>
  <property>
    <name>hbase.regionserver.regionSplitLimit</name>
    <value>2147483647</value>
    <description>Limit for the number of regions after which no more region
    splitting should take place. This is not a hard limit for the number of
    regions but acts as a guideline for the regionserver to stop splitting after
    a certain limit. Default is MAX_INT; i.e. do not block splitting.</description>
  </property>
  <property>
    <name>hbase.regionserver.logroll.period</name>
    <value>3600000</value>
    <description>Period at which we will roll the commit log regardless
    of how many edits it has.</description>
  </property>
  <property>
    <name>hbase.regionserver.logroll.errors.tolerated</name>
    <value>2</value>
    <description>The number of consecutive WAL close errors we will allow
    before triggering a server abort.  A setting of 0 will cause the
    region server to abort if closing the current WAL writer fails during
    log rolling.  Even a small value (2 or 3) will allow a region server
    to ride over transient HDFS errors.</description>
  </property>
  <property>
    <name>hbase.regionserver.hlog.reader.impl</name>
    <value>org.apache.hadoop.hbase.regionserver.wal.ProtobufLogReader</value>
    <description>The HLog file reader implementation.</description>
  </property>
  <property>
    <name>hbase.regionserver.hlog.writer.impl</name>
    <value>org.apache.hadoop.hbase.regionserver.wal.ProtobufLogWriter</value>
    <description>The HLog file writer implementation.</description>
  </property>
  <property>
    <name>hbase.regionserver.global.memstore.upperLimit</name>
    <value>0.4</value>
    <description>Maximum size of all memstores in a region server before new
      updates are blocked and flushes are forced. Defaults to 40% of heap.
      Updates are blocked and flushes are forced until size of all memstores
      in a region server hits hbase.regionserver.global.memstore.lowerLimit.</description>
  </property>
  <property>
    <name>hbase.regionserver.global.memstore.lowerLimit</name>
    <value>0.38</value>
    <description>Maximum size of all memstores in a region server before
      flushes are forced. Defaults to 38% of heap.
      This value equal to hbase.regionserver.global.memstore.upperLimit causes
      the minimum possible flushing to occur when updates are blocked due to
      memstore limiting.</description>
  </property>
  <property>
    <name>hbase.regionserver.optionalcacheflushinterval</name>
    <value>3600000</value>
    <description>
    Maximum amount of time an edit lives in memory before being automatically flushed.
    Default 1 hour. Set it to 0 to disable automatic flushing.</description>
  </property>
  <property>
    <name>hbase.regionserver.catalog.timeout</name>
    <value>600000</value>
    <description>Timeout value for the Catalog Janitor from the regionserver to META.</description>
  </property>
  <property>
    <name>hbase.regionserver.dns.interface</name>
    <value>default</value>
    <description>The name of the Network Interface from which a region server
      should report its IP address.</description>
  </property>
  <property>
    <name>hbase.regionserver.dns.nameserver</name>
    <value>default</value>
    <description>The host name or IP address of the name server (DNS)
      which a region server should use to determine the host name used by the
      master for communication and display purposes.</description>
  </property>

  <!--ZooKeeper configuration-->
  <property>
    <name>zookeeper.session.timeout</name>
    <value>90000</value>
    <description>ZooKeeper session timeout in milliseconds. It is used in two different ways.
      First, this value is used in the ZK client that HBase uses to connect to the ensemble.
      It is also used by HBase when it starts a ZK server and it is passed as the 'maxSessionTimeout'. See
      http://hadoop.apache.org/zookeeper/docs/current/zookeeperProgrammers.html#ch_zkSessions.
      For example, if a HBase region server connects to a ZK ensemble that's also managed by HBase, then the
      session timeout will be the one specified by this configuration. But, a region server that connects
      to an ensemble managed with a different configuration will be subjected that ensemble's maxSessionTimeout. So,
      even though HBase might propose using 90 seconds, the ensemble can have a max timeout lower than this and
      it will take precedence. The current default that ZK ships with is 40 seconds, which is lower than HBase's.
    </description>
  </property>
  <property>
    <name>zookeeper.znode.parent</name>
    <value>/hbase</value>
    <description>Root ZNode for HBase in ZooKeeper. All of HBase's ZooKeeper
      files that are configured with a relative path will go under this node.
      By default, all of HBase's ZooKeeper file path are configured with a
      relative path, so they will all go under this directory unless changed.</description>
  </property>
  <property>
    <name>zookeeper.znode.rootserver</name>
    <value>root-region-server</value>
    <description>Path to ZNode holding root region location. This is written by
      the master and read by clients and region servers. If a relative path is
      given, the parent folder will be ${zookeeper.znode.parent}. By default,
      this means the root location is stored at /hbase/root-region-server.</description>
  </property>
  <property>
    <name>zookeeper.znode.acl.parent</name>
    <value>acl</value>
    <description>Root ZNode for access control lists.</description>
  </property>
  <property>
    <name>hbase.zookeeper.dns.interface</name>
    <value>default</value>
    <description>The name of the Network Interface from which a ZooKeeper server
      should report its IP address.</description>
  </property>
  <property>
    <name>hbase.zookeeper.dns.nameserver</name>
    <value>default</value>
    <description>The host name or IP address of the name server (DNS)
      which a ZooKeeper server should use to determine the host name used by the
      master for communication and display purposes.</description>
  </property>
  <!--
  The following three properties are used together to create the list of
  host:peer_port:leader_port quorum servers for ZooKeeper.
  -->
  <property>
    <name>hbase.zookeeper.peerport</name>
    <value>2888</value>
    <description>Port used by ZooKeeper peers to talk to each other.
    Seehttp://hadoop.apache.org/zookeeper/docs/r3.1.1/zookeeperStarted.html#sc_RunningReplicatedZooKeeper
    for more information.</description>
  </property>
  <property>
    <name>hbase.zookeeper.leaderport</name>
    <value>3888</value>
    <description>Port used by ZooKeeper for leader election.
    See http://hadoop.apache.org/zookeeper/docs/r3.1.1/zookeeperStarted.html#sc_RunningReplicatedZooKeeper
    for more information.</description>
  </property>
  <!-- End of properties used to generate ZooKeeper host:port quorum list. -->
  <property>
    <name>hbase.zookeeper.useMulti</name>
    <value>false</value>
    <description>Instructs HBase to make use of ZooKeeper's multi-update functionality.
    This allows certain ZooKeeper operations to complete more quickly and prevents some issues
    with rare Replication failure scenarios (see the release note of HBASE-2611 for an example).
    IMPORTANT: only set this to true if all ZooKeeper servers in the cluster are on version 3.4+
    and will not be downgraded.  ZooKeeper versions before 3.4 do not support multi-update and
    will not fail gracefully if multi-update is invoked (see ZOOKEEPER-1495).</description>
  </property>
  <property>
    <name>hbase.config.read.zookeeper.config</name>
    <value>false</value>
    <description>
        Set to true to allow HBaseConfiguration to read the
        zoo.cfg file for ZooKeeper properties. Switching this to true
        is not recommended, since the functionality of reading ZK
        properties from a zoo.cfg file has been deprecated.</description>
  </property>
  <!--
  Beginning of properties that are directly mapped from ZooKeeper's zoo.cfg.
  All properties with an "hbase.zookeeper.property." prefix are converted for
  ZooKeeper's configuration. Hence, if you want to add an option from zoo.cfg,
  e.g.  "initLimit=10" you would append the following to your configuration:
    <property>
      <name>hbase.zookeeper.property.initLimit</name>
      <value>10</value>
    </property>
  -->
  <property>
    <name>hbase.zookeeper.property.initLimit</name>
    <value>10</value>
    <description>Property from ZooKeeper's config zoo.cfg.
    The number of ticks that the initial synchronization phase can take.</description>
  </property>
  <property>
    <name>hbase.zookeeper.property.syncLimit</name>
    <value>5</value>
    <description>Property from ZooKeeper's config zoo.cfg.
    The number of ticks that can pass between sending a request and getting an
    acknowledgment.</description>
  </property>
  <property>
    <name>hbase.zookeeper.property.dataDir</name>
    <value>${hbase.tmp.dir}/zookeeper</value>
    <description>Property from ZooKeeper's config zoo.cfg.
    The directory where the snapshot is stored.</description>
  </property>
  <property>
    <name>hbase.zookeeper.property.clientPort</name>
    <value>2181</value>
    <description>Property from ZooKeeper's config zoo.cfg.
    The port at which the clients will connect.</description>
  </property>
  <property>
    <name>hbase.zookeeper.property.maxClientCnxns</name>
    <value>300</value>
    <description>Property from ZooKeeper's config zoo.cfg.
    Limit on number of concurrent connections (at the socket level) that a
    single client, identified by IP address, may make to a single member of
    the ZooKeeper ensemble. Set high to avoid zk connection issues running
    standalone and pseudo-distributed.</description>
  </property>
  <!-- End of properties that are directly mapped from ZooKeeper's zoo.cfg -->

  <!--Client configurations-->
  <property>
    <name>hbase.client.write.buffer</name>
    <value>2097152</value>
    <description>Default size of the HTable client write buffer in bytes.
    A bigger buffer takes more memory -- on both the client and server
    side since server instantiates the passed write buffer to process
    it -- but a larger buffer size reduces the number of RPCs made.
    For an estimate of server-side memory-used, evaluate
    hbase.client.write.buffer * hbase.regionserver.handler.count</description>
  </property>
  <property>
    <name>hbase.client.pause</name>
    <value>100</value>
    <description>General client pause value.  Used mostly as value to wait
    before running a retry of a failed get, region lookup, etc.
    See hbase.client.retries.number for description of how we backoff from
    this initial pause amount and how this pause works w/ retries.</description>
  </property>
  <property>
    <name>hbase.client.retries.number</name>
    <value>35</value>
    <description>Maximum retries.  Used as maximum for all retryable
    operations such as the getting of a cell's value, starting a row update,
    etc.  Retry interval is a rough function based on hbase.client.pause.  At
    first we retry at this interval but then with backoff, we pretty quickly reach
    retrying every ten seconds.  See HConstants#RETRY_BACKOFF for how the backup
    ramps up.  Change this setting and hbase.client.pause to suit your workload.</description>
  </property>
  <property>
    <name>hbase.client.scanner.caching</name>
    <value>100</value>
    <description>Number of rows that will be fetched when calling next
    on a scanner if it is not served from (local, client) memory. Higher
    caching values will enable faster scanners but will eat up more memory
    and some calls of next may take longer and longer times when the cache is empty.
    Do not set this value such that the time between invocations is greater
    than the scanner timeout; i.e. hbase.client.scanner.timeout.period</description>
  </property>
  <property>
    <name>hbase.client.keyvalue.maxsize</name>
    <value>10485760</value>
    <description>Specifies the combined maximum allowed size of a KeyValue
    instance. This is to set an upper boundary for a single entry saved in a
    storage file. Since they cannot be split it helps avoiding that a region
    cannot be split any further because the data is too large. It seems wise
    to set this to a fraction of the maximum region size. Setting it to zero
    or less disables the check.</description>
  </property>
  <property>
    <name>hbase.client.scanner.timeout.period</name>
    <value>60000</value>
    <description>Client scanner lease period in milliseconds.</description>
  </property>
  <property>
    <name>hbase.client.localityCheck.threadPoolSize</name>
    <value>2</value>
  </property>

  <!--Miscellaneous configuration-->
  <property>
    <name>hbase.bulkload.retries.number</name>
    <value>0</value>
    <description>Maximum retries.  This is maximum number of iterations
    to atomic bulk loads are attempted in the face of splitting operations
    0 means never give up.</description>
  </property>
  <property>
    <name>hbase.balancer.period
    </name>
    <value>300000</value>
    <description>Period at which the region balancer runs in the Master.</description>
  </property>
  <property>
    <name>hbase.regions.slop</name>
    <value>0.2</value>
    <description>Rebalance if any regionserver has average + (average * slop) regions.</description>
  </property>
  <property>
    <name>hbase.server.thread.wakefrequency</name>
    <value>10000</value>
    <description>Time to sleep in between searches for work (in milliseconds).
    Used as sleep interval by service threads such as log roller.</description>
  </property>
  <property>
    <name>hbase.server.versionfile.writeattempts</name>
    <value>3</value>
    <description>
    How many time to retry attempting to write a version file
    before just aborting. Each attempt is seperated by the
    hbase.server.thread.wakefrequency milliseconds.</description>
  </property>
  <property>
    <name>hbase.hregion.memstore.flush.size</name>
    <value>134217728</value>
    <description>
    Memstore will be flushed to disk if size of the memstore
    exceeds this number of bytes.  Value is checked by a thread that runs
    every hbase.server.thread.wakefrequency.</description>
  </property>
  <property>
    <name>hbase.hregion.preclose.flush.size</name>
    <value>5242880</value>
    <description>
      If the memstores in a region are this size or larger when we go
      to close, run a "pre-flush" to clear out memstores before we put up
      the region closed flag and take the region offline.  On close,
      a flush is run under the close flag to empty memory.  During
      this time the region is offline and we are not taking on any writes.
      If the memstore content is large, this flush could take a long time to
      complete.  The preflush is meant to clean out the bulk of the memstore
      before putting up the close flag and taking the region offline so the
      flush that runs under the close flag has little to do.</description>
  </property>
  <property>
    <name>hbase.hregion.memstore.block.multiplier</name>
    <value>2</value>
    <description>
    Block updates if memstore has hbase.hregion.block.memstore
    time hbase.hregion.flush.size bytes.  Useful preventing
    runaway memstore during spikes in update traffic.  Without an
    upper-bound, memstore fills such that when it flushes the
    resultant flush files take a long time to compact or split, or
    worse, we OOME.</description>
  </property>
  <property>
    <name>hbase.hregion.memstore.mslab.enabled</name>
    <value>true</value>
    <description>
      Enables the MemStore-Local Allocation Buffer,
      a feature which works to prevent heap fragmentation under
      heavy write loads. This can reduce the frequency of stop-the-world
      GC pauses on large heaps.</description>
  </property>
  <property>
    <name>hbase.hregion.max.filesize</name>
    <value>10737418240</value>
    <description>
    Maximum HStoreFile size. If any one of a column families' HStoreFiles has
    grown to exceed this value, the hosting HRegion is split in two.</description>
  </property>
  <property>
    <name>hbase.hregion.majorcompaction</name>
    <value>604800000</value>
    <description>The time (in miliseconds) between 'major' compactions of all
    HStoreFiles in a region.  Default: Set to 7 days.  Major compactions tend to
    happen exactly when you need them least so enable them such that they run at
    off-peak for your deploy; or, since this setting is on a periodicity that is
    unlikely to match your loading, run the compactions via an external
    invocation out of a cron job or some such.</description>
  </property>
  <property>
    <name>hbase.hregion.majorcompaction.jitter</name>
    <value>0.50</value>
    <description>Jitter outer bound for major compactions.
    On each regionserver, we multiply the hbase.region.majorcompaction
    interval by some random fraction that is inside the bounds of this
    maximum.  We then add this + or - product to when the next
    major compaction is to run.  The idea is that major compaction
    does happen on every regionserver at exactly the same time.  The
    smaller this number, the closer the compactions come together.</description>
  </property>
  <property>
    <name>hbase.hstore.compactionThreshold</name>
    <value>3</value>
    <description>
    If more than this number of HStoreFiles in any one HStore
    (one HStoreFile is written per flush of memstore) then a compaction
    is run to rewrite all HStoreFiles files as one.  Larger numbers
    put off compaction but when it runs, it takes longer to complete.</description>
  </property>
  <property>
    <name>hbase.hstore.blockingStoreFiles</name>
    <value>10</value>
    <description>
    If more than this number of StoreFiles in any one Store
    (one StoreFile is written per flush of MemStore) then updates are
    blocked for this HRegion until a compaction is completed, or
    until hbase.hstore.blockingWaitTime has been exceeded.</description>
  </property>
  <property>
    <name>hbase.hstore.blockingWaitTime</name>
    <value>90000</value>
    <description>
    The time an HRegion will block updates for after hitting the StoreFile
    limit defined by hbase.hstore.blockingStoreFiles.
    After this time has elapsed, the HRegion will stop blocking updates even
    if a compaction has not been completed.</description>
  </property>
  <property>
    <name>hbase.hstore.compaction.max</name>
    <value>10</value>
    <description>Max number of HStoreFiles to compact per 'minor' compaction.</description>
  </property>
  <property>
    <name>hbase.hstore.compaction.kv.max</name>
    <value>10</value>
    <description>How many KeyValues to read and then write in a batch when flushing
        or compacting.  Do less if big KeyValues and problems with OOME.
        Do more if wide, small rows.</description>
  </property>
  <property>
    <name>hbase.storescanner.parallel.seek.enable</name>
    <value>false</value>
    <description>
      Enables StoreFileScanner parallel-seeking in StoreScanner,
      a feature which can reduce response latency under special conditions.</description>
  </property>
  <property>
    <name>hbase.storescanner.parallel.seek.threads</name>
    <value>10</value>
    <description>
      The default thread pool size if parallel-seeking feature enabled.</description>
  </property>
  <property>
    <name>hfile.block.cache.size</name>
    <value>0.4</value>
    <description>Percentage of maximum heap (-Xmx setting) to allocate to block cache
        used by HFile/StoreFile. Default of 0.4 means allocate 40%.
        Set to 0 to disable but it's not recommended; you need at least
        enough cache to hold the storefile indices.</description>
  </property>
  <property>
      <name>hfile.block.index.cacheonwrite</name>
      <value>false</value>
      <description>This allows to put non-root multi-level index blocks into the block
          cache at the time the index is being written.</description>
  </property>
  <property>
      <name>hfile.index.block.max.size</name>
      <value>131072</value>
      <description>When the size of a leaf-level, intermediate-level, or root-level
          index block in a multi-level block index grows to this size, the
          block is written out and a new block is started.</description>
  </property>
  <property>
      <name>hfile.format.version</name>
      <value>2</value>
      <description>The HFile format version to use for new files. Set this to 1 to test
          backwards-compatibility. The default value of this option should be
          consistent with FixedFileTrailer.MAX_VERSION.</description>
  </property>
  <property>
      <name>hfile.block.bloom.cacheonwrite</name>
      <value>false</value>
      <description>Enables cache-on-write for inline blocks of a compound Bloom filter.</description>
  </property>
  <property>
      <name>io.storefile.bloom.block.size</name>
      <value>131072</value>
      <description>The size in bytes of a single block ("chunk") of a compound Bloom
          filter. This size is approximate, because Bloom blocks can only be
          inserted at data block boundaries, and the number of keys per data
          block varies.</description>
  </property>
  <property>
      <name>hbase.rs.cacheblocksonwrite</name>
      <value>false</value>
      <description>Whether an HFile block should be added to the block cache when the
          block is finished.</description>
  </property>
  <property>
    <name>hbase.rpc.server.engine</name>
    <value>org.apache.hadoop.hbase.ipc.ProtobufRpcServerEngine</value>
    <description>Implementation of org.apache.hadoop.hbase.ipc.RpcServerEngine to be
    used for server RPC call marshalling.</description>
  </property>
  <property>
    <name>hbase.rpc.timeout</name>
    <value>60000</value>
    <description>This is for the RPC layer to define how long HBase client applications
        take for a remote call to time out. It uses pings to check connections
        but will eventually throw a TimeoutException.</description>
  </property>
  <property>
    <name>hbase.rpc.shortoperation.timeout</name>
    <value>10000</value>
    <description>This is another version of "hbase.rpc.timeout". For those RPC operation
        within cluster, we rely on this configuration to set a short timeout limitation
        for short operation. For example, short rpc timeout for region server's trying
        to report to active master can benefit quicker master failover process.</description>
  </property>
  <property>
    <name>hbase.ipc.client.tcpnodelay</name>
    <value>true</value>
    <description>Set no delay on rpc socket connections.  See
    http://docs.oracle.com/javase/1.5.0/docs/api/java/net/Socket.html#getTcpNoDelay()</description>
  </property>
  <!-- The following properties configure authentication information for
       HBase processes when using Kerberos security.  There are no default
       values, included here for documentation purposes -->
  <property>
    <name>hbase.master.keytab.file</name>
    <value></value>
    <description>Full path to the kerberos keytab file to use for logging in
    the configured HMaster server principal.</description>
  </property>
  <property>
    <name>hbase.master.kerberos.principal</name>
    <value></value>
    <description>Ex. "hbase/_HOST@EXAMPLE.COM".  The kerberos principal name
    that should be used to run the HMaster process.  The principal name should
    be in the form: user/hostname@DOMAIN.  If "_HOST" is used as the hostname
    portion, it will be replaced with the actual hostname of the running
    instance.</description>
  </property>
  <property>
    <name>hbase.regionserver.keytab.file</name>
    <value></value>
    <description>Full path to the kerberos keytab file to use for logging in
    the configured HRegionServer server principal.</description>
  </property>
  <property>
    <name>hbase.regionserver.kerberos.principal</name>
    <value></value>
    <description>Ex. "hbase/_HOST@EXAMPLE.COM".  The kerberos principal name
    that should be used to run the HRegionServer process.  The principal name
    should be in the form: user/hostname@DOMAIN.  If "_HOST" is used as the
    hostname portion, it will be replaced with the actual hostname of the
    running instance.  An entry for this principal must exist in the file
    specified in hbase.regionserver.keytab.file</description>
  </property>
  <!-- Additional configuration specific to HBase security -->
  <property>
    <name>hadoop.policy.file</name>
    <value>hbase-policy.xml</value>
    <description>The policy configuration file used by RPC servers to make
      authorization decisions on client requests.  Only used when HBase
      security is enabled.</description>
  </property>
  <property>
    <name>hbase.superuser</name>
    <value></value>
    <description>List of users or groups (comma-separated), who are allowed
    full privileges, regardless of stored ACLs, across the cluster.
    Only used when HBase security is enabled.</description>
  </property>
  <property>
    <name>hbase.auth.key.update.interval</name>
    <value>86400000</value>
    <description>The update interval for master key for authentication tokens
    in servers in milliseconds.  Only used when HBase security is enabled.</description>
  </property>
  <property>
    <name>hbase.auth.token.max.lifetime</name>
    <value>604800000</value>
    <description>The maximum lifetime in milliseconds after which an
    authentication token expires.  Only used when HBase security is enabled.</description>
  </property>
  <property>
    <name>hbase.ipc.client.fallback-to-simple-auth-allowed</name>
    <value>false</value>
    <description>When a client is configured to attempt a secure connection, but attempts to
      connect to an insecure server, that server may instruct the client to
      switch to SASL SIMPLE (unsecure) authentication. This setting controls
      whether or not the client will accept this instruction from the server.
      When false (the default), the client will not allow the fallback to SIMPLE
      authentication, and will abort the connection.</description>
  </property>
  <property>
    <name>hbase.coprocessor.region.classes</name>
    <value></value>
    <description>A comma-separated list of Coprocessors that are loaded by
    default on all tables. For any override coprocessor method, these classes
    will be called in order. After implementing your own Coprocessor, just put
    it in HBase's classpath and add the fully qualified class name here.
    A coprocessor can also be loaded on demand by setting HTableDescriptor.</description>
  </property>
  <property>
    <name>hbase.rest.port</name>
    <value>8080</value>
    <description>The port for the HBase REST server.</description>
  </property>
  <property>
    <name>hbase.rest.readonly</name>
    <value>false</value>
    <description>Defines the mode the REST server will be started in. Possible values are:
    false: All HTTP methods are permitted - GET/PUT/POST/DELETE.
    true: Only the GET method is permitted.</description>
  </property>
  <property>
    <name>hbase.rest.threads.max</name>
    <value>100</value>
    <description>The maximum number of threads of the REST server thread pool.
        Threads in the pool are reused to process REST requests. This
        controls the maximum number of requests processed concurrently.
        It may help to control the memory used by the REST server to
        avoid OOM issues. If the thread pool is full, incoming requests
        will be queued up and wait for some free threads.</description>
  </property>
  <property>
    <name>hbase.rest.threads.min</name>
    <value>2</value>
    <description>The minimum number of threads of the REST server thread pool.
        The thread pool always has at least these number of threads so
        the REST server is ready to serve incoming requests.</description>
  </property>
  <property skipInDoc="true">
    <name>hbase.defaults.for.version</name>
    <value>@@@VERSION@@@</value>
    <description>This defaults file was compiled for version ${project.version}. This variable is used
    to make sure that a user doesn't have an old version of hbase-default.xml on the
    classpath.</description>
  </property>
  <property>
    <name>hbase.defaults.for.version.skip</name>
    <value>false</value>
    <description>Set to true to skip the 'hbase.defaults.for.version' check.
    Setting this to true can be useful in contexts other than
    the other side of a maven generation; i.e. running in an
    ide.  You'll want to set this boolean to true to avoid
    seeing the RuntimException complaint: "hbase-default.xml file
    seems to be for and old version of HBase (\${hbase.version}), this
    version is X.X.X-SNAPSHOT"</description>
  </property>
  <property>
    <name>hbase.coprocessor.master.classes</name>
    <value></value>
    <description>A comma-separated list of
    org.apache.hadoop.hbase.coprocessor.MasterObserver coprocessors that are
    loaded by default on the active HMaster process. For any implemented
    coprocessor methods, the listed classes will be called in order. After
    implementing your own MasterObserver, just put it in HBase's classpath
    and add the fully qualified class name here.</description>
  </property>
  <property>
      <name>hbase.coprocessor.abortonerror</name>
      <value>false</value>
      <description>Set to true to cause the hosting server (master or regionserver) to
      abort if a coprocessor throws a Throwable object that is not IOException or
      a subclass of IOException. Setting it to true might be useful in development
      environments where one wants to terminate the server as soon as possible to
      simplify coprocessor failure analysis.</description>
  </property>
  <property>
    <name>hbase.online.schema.update.enable</name>
    <value>false</value>
    <description>Set true to enable online schema changes.</description>
  </property>
  <property>
    <name>hbase.table.lock.enable</name>
    <value>true</value>
    <description>Set to true to enable locking the table in zookeeper for schema change operations.
    Table locking from master prevents concurrent schema modifications to corrupt table
    state.</description>
  </property>
  <property>
    <name>hbase.thrift.minWorkerThreads</name>
    <value>16</value>
    <description>The "core size" of the thread pool. New threads are created on every
    connection until this many threads are created.</description>
  </property>
  <property>
    <name>hbase.thrift.maxWorkerThreads</name>
    <value>1000</value>
    <description>The maximum size of the thread pool. When the pending request queue
    overflows, new threads are created until their number reaches this number.
    After that, the server starts dropping connections.</description>
  </property>
  <property>
    <name>hbase.thrift.maxQueuedRequests</name>
    <value>1000</value>
    <description>The maximum number of pending Thrift connections waiting in the queue. If
     there are no idle threads in the pool, the server queues requests. Only
     when the queue overflows, new threads are added, up to
     hbase.thrift.maxQueuedRequests threads.</description>
  </property>
  <property>
    <name>hbase.thrift.htablepool.size.max</name>
    <value>1000</value>
    <description>The upper bound for the table pool used in the Thrift gateways server.
      Since this is per table name, we assume a single table and so with 1000 default
      worker threads max this is set to a matching number. For other workloads this number
      can be adjusted as needed.
    </description>
  </property>
  <property>
    <name>hbase.offheapcache.percentage</name>
    <value>0</value>
    <description>The amount of off heap space to be allocated towards the experimental
     off heap cache. If you desire the cache to be disabled, simply set this
     value to 0.</description>
  </property>
  <property>
    <name>hbase.data.umask.enable</name>
    <value>false</value>
    <description>Enable, if true, that file permissions should be assigned
      to the files written by the regionserver</description>
  </property>
  <property>
    <name>hbase.data.umask</name>
    <value>000</value>
    <description>File permissions that should be used to write data
      files when hbase.data.umask.enable is true</description>
  </property>
  <property>
    <name>hbase.metrics.showTableName</name>
    <value>true</value>
    <description>Whether to include the prefix "tbl.tablename" in per-column family metrics.
        If true, for each metric M, per-cf metrics will be reported for tbl.T.cf.CF.M, if false,
        per-cf metrics will be aggregated by column-family across tables, and reported for cf.CF.M.
        In both cases, the aggregated metric M across tables and cfs will be reported.</description>
  </property>
  <property>
    <name>hbase.metrics.exposeOperationTimes</name>
    <value>true</value>
    <description>Whether to report metrics about time taken performing an
      operation on the region server.  Get, Put, Delete, Increment, and Append can all
      have their times exposed through Hadoop metrics per CF and per region.</description>
  </property>
  <property>
    <name>hbase.snapshot.enabled</name>
    <value>true</value>
    <description>Set to true to allow snapshots to be taken / restored / cloned.</description>
  </property>
  <property>
    <name>hbase.server.compactchecker.interval.multiplier</name>
    <value>1000</value>
    <description>The number that determines how often we scan to see if compaction is necessary.
        Normally, compactions are done after some events (such as memstore flush), but if
        region didn't receive a lot of writes for some time, or due to different compaction
        policies, it may be necessary to check it periodically. The interval between checks is
        hbase.server.compactchecker.interval.multiplier multiplied by
        hbase.server.thread.wakefrequency.</description>
  </property>
  <property>
    <name>hbase.lease.recovery.timeout</name>
    <value>900000</value>
    <description>How long we wait on dfs lease recovery in total before giving up.</description>
  </property>
  <property>
    <name>hbase.lease.recovery.dfs.timeout</name>
    <value>64000</value>
    <description>How long between dfs recover lease invocations. Should be larger than the sum of
        the time it takes for the namenode to issue a block recovery command as part of
        datanode; dfs.heartbeat.interval and the time it takes for the primary
        datanode, performing block recovery to timeout on a dead datanode; usually
        dfs.socket.timeout. See the end of HBASE-8389 for more.</description>
  </property>
  <property>
    <name>hbase.regionserver.checksum.verify</name>
    <value>true</value>
    <description>
        If set to true, HBase will read data and then verify checksums for
        hfile blocks. Checksum verification inside HDFS will be switched off.
        If the hbase-checksum verification fails, then it will switch back to
        using HDFS checksums.
    </description>
  </property>
  <property>
    <name>hbase.hstore.bytes.per.checksum</name>
    <value>16384</value>
    <description>
        Number of bytes in a newly created checksum chunk for HBase-level
        checksums in hfile blocks.
    </description>
  </property>
  <property>
    <name>hbase.hstore.checksum.algorithm</name>
    <value>CRC32</value>
    <description>
      Name of an algorithm that is used to compute checksums. Possible values
      are NULL, CRC32, CRC32C.
    </description>
  </property>

  <property>
    <name>hbase.status.published</name>
    <value>false</value>
    <description>
      This setting activates the publication by the master of the status of the region server.
      When a region server dies and its recovery starts, the master will push this information
      to the client application, to let them cut the connection immediately instead of waiting
      for a timeout.
    </description>
  </property>
  <property>
    <name>hbase.status.publisher.class</name>
    <value>org.apache.hadoop.hbase.master.ClusterStatusPublisher$MulticastPublisher</value>
    <description>
      Implementation of the status publication with a multicast message.
    </description>
  </property>
  <property>
    <name>hbase.status.listener.class</name>
    <value>org.apache.hadoop.hbase.client.ClusterStatusListener$MulticastListener</value>
    <description>
      Implementation of the status listener with a multicast message.
    </description>
  </property>
  <property>
    <name>hbase.status.multicast.address.ip</name>
    <value>226.1.1.3</value>
    <description>
      Multicast address to use for the status publication by multicast.
    </description>
  </property>
  <property>
    <name>hbase.status.multicast.address.port</name>
    <value>6100</value>
    <description>
      Multicast port to use for the status publication by multicast.
    </description>
  </property>

  <property>
    <name>hbase.dynamic.jars.dir</name>
    <value>${hbase.rootdir}/lib</value>
    <description>
      The directory from which the custom filter/co-processor jars can be loaded
      dynamically by the region server without the need to restart. However,
      an already loaded filter/co-processor class would not be un-loaded. See
      HBASE-1936 for more details.
    </description>
  </property>
  <property>
    <name>hbase.security.authentication</name>
    <value>simple</value>
    <description>
      Controls whether or not secure authentication is enabled for HBase.
      Possible values are 'simple' (no authentication), and 'kerberos'.
    </description>
  </property>
  <property>
    <name>hbase.rest.filter.classes</name>
    <value>org.apache.hadoop.hbase.rest.filter.GzipFilter</value>
    <description>
      Servlet filters for REST service.
    </description>
  </property>
  <property>
    <name>hbase.master.loadbalancer.class</name>
    <value>org.apache.hadoop.hbase.master.balancer.StochasticLoadBalancer</value>
    <description>
      Class used to execute the regions balancing when the period occurs.
      See the class comment for more on how it works
      http://hbase.apache.org/devapidocs/org/apache/hadoop/hbase/master/balancer/StochasticLoadBalancer.html
      It replaces the DefaultLoadBalancer as the default (since renamed
      as the SimpleLoadBalancer).
    </description>
  </property>
</configuration>
  评论这张
 
阅读(882)| 评论(0)
推荐 转载

历史上的今天

在LOFTER的更多文章

评论

<#--最新日志,群博日志--> <#--推荐日志--> <#--引用记录--> <#--博主推荐--> <#--随机阅读--> <#--首页推荐--> <#--历史上的今天--> <#--被推荐日志--> <#--上一篇,下一篇--> <#-- 热度 --> <#-- 网易新闻广告 --> <#--右边模块结构--> <#--评论模块结构--> <#--引用模块结构--> <#--博主发起的投票-->
 
 
 
 
 
 
 
 
 
 
 
 
 
 

页脚

网易公司版权所有 ©1997-2017