You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by nd...@apache.org on 2015/08/19 00:36:00 UTC
[08/15] hbase git commit: HBASE-14066 clean out old docbook docs from
branch-1.
http://git-wip-us.apache.org/repos/asf/hbase/blob/0acbff24/src/main/docbkx/ops_mgt.xml
----------------------------------------------------------------------
diff --git a/src/main/docbkx/ops_mgt.xml b/src/main/docbkx/ops_mgt.xml
deleted file mode 100644
index 0af8f02..0000000
--- a/src/main/docbkx/ops_mgt.xml
+++ /dev/null
@@ -1,2488 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<chapter
- version="5.0"
- xml:id="ops_mgt"
- xmlns="http://docbook.org/ns/docbook"
- xmlns:xlink="http://www.w3.org/1999/xlink"
- xmlns:xi="http://www.w3.org/2001/XInclude"
- xmlns:svg="http://www.w3.org/2000/svg"
- xmlns:m="http://www.w3.org/1998/Math/MathML"
- xmlns:html="http://www.w3.org/1999/xhtml"
- xmlns:db="http://docbook.org/ns/docbook">
- <!--
-/**
- * Licensed to the Apache Software Foundation (ASF) under one
- * or more contributor license agreements. See the NOTICE file
- * distributed with this work forf 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.
- */
--->
- <title>Apache HBase Operational Management</title>
- <para> This chapter will cover operational tools and practices required of a running Apache HBase
- cluster. The subject of operations is related to the topics of <xref
- linkend="trouble" />, <xref
- linkend="performance" />, and <xref
- linkend="configuration" /> but is a distinct topic in itself. </para>
-
- <section
- xml:id="tools">
- <title>HBase Tools and Utilities</title>
-
- <para>HBase provides several tools for administration, analysis, and debugging of your cluster.
- The entry-point to most of these tools is the <filename>bin/hbase</filename> command, though
- some tools are available in the <filename>dev-support/</filename> directory.</para>
- <para>To see usage instructions for <filename>bin/hbase</filename> command, run it with no
- arguments, or with the <option>-h</option> argument. These are the usage instructions for
- HBase 0.98.x. Some commands, such as <command>version</command>, <command>pe</command>,
- <command>ltt</command>, <command>clean</command>, are not available in previous
- versions.</para>
- <screen>
-$ <userinput>bin/hbase</userinput>
-<![CDATA[Usage: hbase [<options>] <command> [<args>]]]>
-Options:
- --config DIR Configuration direction to use. Default: ./conf
- --hosts HOSTS Override the list in 'regionservers' file
-
-Commands:
-Some commands take arguments. Pass no args or -h for usage.
- shell Run the HBase shell
- hbck Run the hbase 'fsck' tool
- hlog Write-ahead-log analyzer
- hfile Store file analyzer
- zkcli Run the ZooKeeper shell
- upgrade Upgrade hbase
- master Run an HBase HMaster node
- regionserver Run an HBase HRegionServer node
- zookeeper Run a Zookeeper server
- rest Run an HBase REST server
- thrift Run the HBase Thrift server
- thrift2 Run the HBase Thrift2 server
- clean Run the HBase clean up script
- classpath Dump hbase CLASSPATH
- mapredcp Dump CLASSPATH entries required by mapreduce
- pe Run PerformanceEvaluation
- ltt Run LoadTestTool
- version Print the version
- CLASSNAME Run the class named CLASSNAME
- </screen>
- <para>Some of the tools and utilities below are Java classes which are passed directly to the
- <filename>bin/hbase</filename> command, as referred to in the last line of the usage
- instructions. Others, such as <command>hbase shell</command> (<xref linkend="shell"/>),
- <command>hbase upgrade</command> (<xref linkend="upgrading"/>), and <command>hbase
- thrift</command> (<xref linkend="thrift"/>), are documented elsewhere in this guide.</para>
- <section
- xml:id="canary">
- <title>Canary</title>
- <para> There is a Canary class can help users to canary-test the HBase cluster status, with
- every column-family for every regions or regionservers granularity. To see the usage, use
- the <literal>--help</literal> parameter. </para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase org.apache.hadoop.hbase.tool.Canary -help
-
-Usage: bin/hbase org.apache.hadoop.hbase.tool.Canary [opts] [table1 [table2]...] | [regionserver1 [regionserver2]..]
- where [opts] are:
- -help Show this help and exit.
- -regionserver replace the table argument to regionserver,
- which means to enable regionserver mode
- -daemon Continuous check at defined intervals.
- -interval <N> Interval between checks (sec)
- -e Use region/regionserver as regular expression
- which means the region/regionserver is regular expression pattern
- -f <B> stop whole program if first error occurs, default is true
- -t <N> timeout for a check, default is 600000 (milliseconds)</screen>
- <para> This tool will return non zero error codes to user for collaborating with other
- monitoring tools, such as Nagios. The error code definitions are: </para>
- <programlisting language="java">private static final int USAGE_EXIT_CODE = 1;
-private static final int INIT_ERROR_EXIT_CODE = 2;
-private static final int TIMEOUT_ERROR_EXIT_CODE = 3;
-private static final int ERROR_EXIT_CODE = 4;</programlisting>
- <para> Here are some examples based on the following given case. There are two HTable called
- test-01 and test-02, they have two column family cf1 and cf2 respectively, and deployed on
- the 3 regionservers. see following table. </para>
-
- <informaltable>
- <tgroup
- cols="3"
- align="center"
- colsep="1"
- rowsep="1">
- <colspec
- colname="regionserver"
- align="center" />
- <colspec
- colname="test-01"
- align="center" />
- <colspec
- colname="test-02"
- align="center" />
- <thead>
- <row>
- <entry>RegionServer</entry>
- <entry>test-01</entry>
- <entry>test-02</entry>
- </row>
- </thead>
- <tbody>
- <row>
- <entry>rs1</entry>
- <entry>r1</entry>
- <entry>r2</entry>
- </row>
- <row>
- <entry>rs2</entry>
- <entry>r2</entry>
- <entry />
- </row>
- <row>
- <entry>rs3</entry>
- <entry>r2</entry>
- <entry>r1</entry>
- </row>
- </tbody>
- </tgroup>
- </informaltable>
- <para> Following are some examples based on the previous given case. </para>
- <section>
- <title>Canary test for every column family (store) of every region of every table</title>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase org.apache.hadoop.hbase.tool.Canary
-
-3/12/09 03:26:32 INFO tool.Canary: read from region test-01,,1386230156732.0e3c7d77ffb6361ea1b996ac1042ca9a. column family cf1 in 2ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-01,,1386230156732.0e3c7d77ffb6361ea1b996ac1042ca9a. column family cf2 in 2ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-01,0004883,1386230156732.87b55e03dfeade00f441125159f8ca87. column family cf1 in 4ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-01,0004883,1386230156732.87b55e03dfeade00f441125159f8ca87. column family cf2 in 1ms
-...
-13/12/09 03:26:32 INFO tool.Canary: read from region test-02,,1386559511167.aa2951a86289281beee480f107bb36ee. column family cf1 in 5ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-02,,1386559511167.aa2951a86289281beee480f107bb36ee. column family cf2 in 3ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-02,0004883,1386559511167.cbda32d5e2e276520712d84eaaa29d84. column family cf1 in 31ms
-13/12/09 03:26:32 INFO tool.Canary: read from region test-02,0004883,1386559511167.cbda32d5e2e276520712d84eaaa29d84. column family cf2 in 8ms
-</screen>
- <para> So you can see, table test-01 has two regions and two column families, so the Canary
- tool will pick 4 small piece of data from 4 (2 region * 2 store) different stores. This is
- a default behavior of the this tool does. </para>
- </section>
-
- <section>
- <title>Canary test for every column family (store) of every region of specific
- table(s)</title>
- <para> You can also test one or more specific tables.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase orghapache.hadoop.hbase.tool.Canary test-01 test-02</screen>
- </section>
-
- <section>
- <title>Canary test with regionserver granularity</title>
- <para> This will pick one small piece of data from each regionserver, and can also put your
- resionserver name as input options for canary-test specific regionservers.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase org.apache.hadoop.hbase.tool.Canary -regionserver
-
-13/12/09 06:05:17 INFO tool.Canary: Read from table:test-01 on region server:rs2 in 72ms
-13/12/09 06:05:17 INFO tool.Canary: Read from table:test-02 on region server:rs3 in 34ms
-13/12/09 06:05:17 INFO tool.Canary: Read from table:test-01 on region server:rs1 in 56ms</screen>
- </section>
- <section>
- <title>Canary test with regular expression pattern</title>
- <para> This will test both table test-01 and test-02.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase orghapache.hadoop.hbase.tool.Canary -e test-0[1-2]</screen>
- </section>
-
- <section>
- <title>Run canary test as daemon mode</title>
- <para> Run repeatedly with interval defined in option -interval whose default value is 6
- seconds. This daemon will stop itself and return non-zero error code if any error occurs,
- due to the default value of option -f is true.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase orghapache.hadoop.hbase.tool.Canary -daemon</screen>
- <para>Run repeatedly with internal 5 seconds and will not stop itself even error occurs in
- the test.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase orghapache.hadoop.hbase.tool.Canary -daemon -interval 50000 -f false</screen>
- </section>
-
- <section>
- <title>Force timeout if canary test stuck</title>
- <para>In some cases, we suffered the request stucked on the regionserver and not response
- back to the client. The regionserver in problem, would also not indicated to be dead by
- Master, which would bring the clients hung. So we provide the timeout option to kill the
- canary test forcefully and return non-zero error code as well. This run sets the timeout
- value to 60 seconds, the default value is 600 seconds.</para>
- <screen language="bourne">$ ${HBASE_HOME}/bin/hbase orghapache.hadoop.hbase.tool.Canary -t 600000</screen>
- </section>
-
- </section>
-
- <section
- xml:id="health.check">
- <title>Health Checker</title>
- <para>You can configure HBase to run a script on a period and if it fails N times
- (configurable), have the server exit. See <link
- xlink:href="">HBASE-7351 Periodic health check script</link> for configurations and
- detail. </para>
- </section>
-
- <section
- xml:id="driver">
- <title>Driver</title>
- <para>Several frequently-accessed utilities are provided as <code>Driver</code> classes, and executed by
- the <filename>bin/hbase</filename> command. These utilities represent MapReduce jobs which
- run on your cluster. They are run in the following way, replacing
- <replaceable>UtilityName</replaceable> with the utility you want to run. This command
- assumes you have set the environment variable <literal>HBASE_HOME</literal> to the directory
- where HBase is unpacked on your server.</para>
- <screen language="bourne">
-${HBASE_HOME}/bin/hbase org.apache.hadoop.hbase.mapreduce.<replaceable>UtilityName</replaceable>
- </screen>
- <para>The following utilities are available:</para>
- <variablelist>
- <varlistentry>
- <term><command>LoadIncrementalHFiles</command></term>
- <listitem><para>Complete a bulk data load.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>CopyTable</command></term>
- <listitem><para>Export a table from the local cluster to a peer cluster.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>Export</command></term>
- <listitem><para>Write table data to HDFS.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>Import</command></term>
- <listitem><para>Import data written by a previous <command>Export</command> operation.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>ImportTsv</command></term>
- <listitem><para>Import data in TSV format.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>RowCounter</command></term>
- <listitem><para>Count rows in an HBase table.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term><command>replication.VerifyReplication</command></term>
- <listitem><para>Compare the data from tables in two different clusters. WARNING: It
- doesn't work for incrementColumnValues'd cells since the timestamp is changed. Note that
- this command is in a different package than the others.</para></listitem>
- </varlistentry>
- </variablelist>
- <para>Each command except <command>RowCounter</command> accepts a single
- <literal>--help</literal> argument to print usage instructions.</para>
- </section>
- <section
- xml:id="hbck">
- <title>HBase <application>hbck</application></title>
- <subtitle>An <command>fsck</command> for your HBase install</subtitle>
- <para>To run <application>hbck</application> against your HBase cluster run <command>$
- ./bin/hbase hbck</command> At the end of the command's output it prints
- <literal>OK</literal> or <literal>INCONSISTENCY</literal>. If your cluster reports
- inconsistencies, pass <command>-details</command> to see more detail emitted. If
- inconsistencies, run <command>hbck</command> a few times because the inconsistency may be
- transient (e.g. cluster is starting up or a region is splitting). Passing
- <command>-fix</command> may correct the inconsistency (This latter is an experimental
- feature). </para>
- <para>For more information, see <xref
- linkend="hbck.in.depth" />. </para>
- </section>
- <section
- xml:id="hfile_tool2">
- <title>HFile Tool</title>
- <para>See <xref
- linkend="hfile_tool" />.</para>
- </section>
- <section
- xml:id="wal_tools">
- <title>WAL Tools</title>
-
- <section
- xml:id="hlog_tool">
- <title><classname>FSHLog</classname> tool</title>
-
- <para>The main method on <classname>FSHLog</classname> offers manual split and dump
- facilities. Pass it WALs or the product of a split, the content of the
- <filename>recovered.edits</filename>. directory.</para>
-
- <para>You can get a textual dump of a WAL file content by doing the following:</para>
- <screen language="bourne"> $ ./bin/hbase org.apache.hadoop.hbase.regionserver.wal.FSHLog --dump hdfs://example.org:8020/hbase/.logs/example.org,60020,1283516293161/10.10.21.10%3A60020.1283973724012 </screen>
- <para>The return code will be non-zero if issues with the file so you can test wholesomeness
- of file by redirecting <varname>STDOUT</varname> to <code>/dev/null</code> and testing the
- program return.</para>
-
- <para>Similarly you can force a split of a log file directory by doing:</para>
- <screen language="bourne"> $ ./bin/hbase org.apache.hadoop.hbase.regionserver.wal.FSHLog --split hdfs://example.org:8020/hbase/.logs/example.org,60020,1283516293161/</screen>
-
- <section
- xml:id="hlog_tool.prettyprint">
- <title>WAL Pretty Printer</title>
- <para>The WAL Pretty Printer is a tool with configurable options to
- print the contents of a WAL. You can invoke it via the hbase cli with the 'wal' command.
- </para>
- <screen langauge="bourne"> $ ./bin/hbase wal hdfs://example.org:8020/hbase/.logs/example.org,60020,1283516293161/10.10.21.10%3A60020.1283973724012</screen>
- <note>
- <title>WAL Printing in older versions of HBase</title>
- <para>Prior to version 2.0, the WAL Pretty Printer was called the
- <classname>HLogPrettyPrinter</classname>, after an internal name for HBase's write
- ahead log. In those versions, you can pring the contents of a WAL using the same
- configuration as above, but with the 'hlog' command.
- </para>
- <screen langauge="bourne"> $ ./bin/hbase hlog hdfs://example.org:8020/hbase/.logs/example.org,60020,1283516293161/10.10.21.10%3A60020.1283973724012</screen>
- </note>
- </section>
-
- </section>
- </section>
- <section
- xml:id="compression.tool">
- <title>Compression Tool</title>
- <para>See <xref
- linkend="compression.test" />.</para>
- </section>
- <section
- xml:id="copytable">
- <title>CopyTable</title>
- <para> CopyTable is a utility that can copy part or of all of a table, either to the same
- cluster or another cluster. The target table must first exist. The usage is as
- follows:</para>
-
- <screen language="bourne">
-$ <userinput>./bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable --help </userinput>
-/bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable --help
-Usage: CopyTable [general options] [--starttime=X] [--endtime=Y] [--new.name=NEW] [--peer.adr=ADR] <tablename>
-
-Options:
- rs.class hbase.regionserver.class of the peer cluster,
- specify if different from current cluster
- rs.impl hbase.regionserver.impl of the peer cluster,
- startrow the start row
- stoprow the stop row
- starttime beginning of the time range (unixtime in millis)
- without endtime means from starttime to forever
- endtime end of the time range. Ignored if no starttime specified.
- versions number of cell versions to copy
- new.name new table's name
- peer.adr Address of the peer cluster given in the format
- hbase.zookeeer.quorum:hbase.zookeeper.client.port:zookeeper.znode.parent
- families comma-separated list of families to copy
- To copy from cf1 to cf2, give sourceCfName:destCfName.
- To keep the same name, just give "cfName"
- all.cells also copy delete markers and deleted cells
-
-Args:
- tablename Name of the table to copy
-
-Examples:
- To copy 'TestTable' to a cluster that uses replication for a 1 hour window:
- $ bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable --starttime=1265875194289 --endtime=1265878794289 --peer.adr=server1,server2,server3:2181:/hbase --families=myOldCf:myNewCf,cf2,cf3 TestTable
-
-For performance consider the following general options:
- It is recommended that you set the following to >=100. A higher value uses more memory but
- decreases the round trip time to the server and may increase performance.
- -Dhbase.client.scanner.caching=100
- The following should always be set to false, to prevent writing data twice, which may produce
- inaccurate results.
- -Dmapred.map.tasks.speculative.execution=false
- </screen>
- <note>
- <title>Scanner Caching</title>
- <para>Caching for the input Scan is configured via <code>hbase.client.scanner.caching</code>
- in the job configuration. </para>
- </note>
- <note>
- <title>Versions</title>
- <para>By default, CopyTable utility only copies the latest version of row cells unless
- <code>--versions=n</code> is explicitly specified in the command. </para>
- </note>
- <para> See Jonathan Hsieh's <link
- xlink:href="http://www.cloudera.com/blog/2012/06/online-hbase-backups-with-copytable-2/">Online
- HBase Backups with CopyTable</link> blog post for more on <command>CopyTable</command>.
- </para>
- </section>
- <section
- xml:id="export">
- <title>Export</title>
- <para>Export is a utility that will dump the contents of table to HDFS in a sequence file.
- Invoke via:</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.Export <tablename> <outputdir> [<versions> [<starttime> [<endtime>]]]
-</screen>
-
- <para>Note: caching for the input Scan is configured via
- <code>hbase.client.scanner.caching</code> in the job configuration. </para>
- </section>
- <section
- xml:id="import">
- <title>Import</title>
- <para>Import is a utility that will load data that has been exported back into HBase. Invoke
- via:</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.Import <tablename> <inputdir>
-</screen>
- <para>To import 0.94 exported files in a 0.96 cluster or onwards, you need to set system
- property "hbase.import.version" when running the import command as below:</para>
- <screen language="bourne">$ bin/hbase -Dhbase.import.version=0.94 org.apache.hadoop.hbase.mapreduce.Import <tablename> <inputdir>
-</screen>
- </section>
- <section
- xml:id="importtsv">
- <title>ImportTsv</title>
- <para>ImportTsv is a utility that will load data in TSV format into HBase. It has two distinct
- usages: loading data from TSV format in HDFS into HBase via Puts, and preparing StoreFiles
- to be loaded via the <code>completebulkload</code>. </para>
- <para>To load data via Puts (i.e., non-bulk loading):</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.ImportTsv -Dimporttsv.columns=a,b,c <tablename> <hdfs-inputdir>
-</screen>
-
- <para>To generate StoreFiles for bulk-loading:</para>
- <programlisting language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.ImportTsv -Dimporttsv.columns=a,b,c -Dimporttsv.bulk.output=hdfs://storefile-outputdir <tablename> <hdfs-data-inputdir>
-</programlisting>
- <para>These generated StoreFiles can be loaded into HBase via <xref
- linkend="completebulkload" />. </para>
- <section
- xml:id="importtsv.options">
- <title>ImportTsv Options</title>
- <para>Running <command>ImportTsv</command> with no arguments prints brief usage
- information:</para>
- <screen>
-Usage: importtsv -Dimporttsv.columns=a,b,c <tablename> <inputdir>
-
-Imports the given input directory of TSV data into the specified table.
-
-The column names of the TSV data must be specified using the -Dimporttsv.columns
-option. This option takes the form of comma-separated column names, where each
-column name is either a simple column family, or a columnfamily:qualifier. The special
-column name HBASE_ROW_KEY is used to designate that this column should be used
-as the row key for each imported record. You must specify exactly one column
-to be the row key, and you must specify a column name for every column that exists in the
-input data.
-
-By default importtsv will load data directly into HBase. To instead generate
-HFiles of data to prepare for a bulk data load, pass the option:
- -Dimporttsv.bulk.output=/path/for/output
- Note: the target table will be created with default column family descriptors if it does not already exist.
-
-Other options that may be specified with -D include:
- -Dimporttsv.skip.bad.lines=false - fail if encountering an invalid line
- '-Dimporttsv.separator=|' - eg separate on pipes instead of tabs
- -Dimporttsv.timestamp=currentTimeAsLong - use the specified timestamp for the import
- -Dimporttsv.mapper.class=my.Mapper - A user-defined Mapper to use instead of org.apache.hadoop.hbase.mapreduce.TsvImporterMapper
- </screen>
- </section>
- <section
- xml:id="importtsv.example">
- <title>ImportTsv Example</title>
- <para>For example, assume that we are loading data into a table called 'datatsv' with a
- ColumnFamily called 'd' with two columns "c1" and "c2". </para>
- <para>Assume that an input file exists as follows:
- <screen>
-row1 c1 c2
-row2 c1 c2
-row3 c1 c2
-row4 c1 c2
-row5 c1 c2
-row6 c1 c2
-row7 c1 c2
-row8 c1 c2
-row9 c1 c2
-row10 c1 c2
- </screen>
- </para>
- <para>For ImportTsv to use this imput file, the command line needs to look like this:</para>
- <screen language="bourne">
- HADOOP_CLASSPATH=`${HBASE_HOME}/bin/hbase classpath` ${HADOOP_HOME}/bin/hadoop jar ${HBASE_HOME}/hbase-server-VERSION.jar importtsv -Dimporttsv.columns=HBASE_ROW_KEY,d:c1,d:c2 -Dimporttsv.bulk.output=hdfs://storefileoutput datatsv hdfs://inputfile
- </screen>
- <para> ... and in this example the first column is the rowkey, which is why the
- HBASE_ROW_KEY is used. The second and third columns in the file will be imported as "d:c1"
- and "d:c2", respectively. </para>
- </section>
- <section
- xml:id="importtsv.warning">
- <title>ImportTsv Warning</title>
- <para>If you have preparing a lot of data for bulk loading, make sure the target HBase table
- is pre-split appropriately. </para>
- </section>
- <section
- xml:id="importtsv.also">
- <title>See Also</title>
- <para>For more information about bulk-loading HFiles into HBase, see <xref
- linkend="arch.bulk.load" /></para>
- </section>
- </section>
-
- <section
- xml:id="completebulkload">
- <title>CompleteBulkLoad</title>
- <para>The <code>completebulkload</code> utility will move generated StoreFiles into an HBase
- table. This utility is often used in conjunction with output from <xref
- linkend="importtsv" />. </para>
- <para>There are two ways to invoke this utility, with explicit classname and via the
- driver:</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles <hdfs://storefileoutput> <tablename>
-</screen>
- <para> .. and via the Driver..</para>
- <screen language="bourne">HADOOP_CLASSPATH=`${HBASE_HOME}/bin/hbase classpath` ${HADOOP_HOME}/bin/hadoop jar ${HBASE_HOME}/hbase-server-VERSION.jar completebulkload <hdfs://storefileoutput> <tablename>
-</screen>
- <section
- xml:id="completebulkload.warning">
- <title>CompleteBulkLoad Warning</title>
- <para>Data generated via MapReduce is often created with file permissions that are not
- compatible with the running HBase process. Assuming you're running HDFS with permissions
- enabled, those permissions will need to be updated before you run CompleteBulkLoad.</para>
- <para>For more information about bulk-loading HFiles into HBase, see <xref
- linkend="arch.bulk.load" />. </para>
- </section>
-
- </section>
- <section
- xml:id="walplayer">
- <title>WALPlayer</title>
- <para>WALPlayer is a utility to replay WAL files into HBase. </para>
- <para>The WAL can be replayed for a set of tables or all tables, and a timerange can be
- provided (in milliseconds). The WAL is filtered to this set of tables. The output can
- optionally be mapped to another set of tables. </para>
- <para>WALPlayer can also generate HFiles for later bulk importing, in that case only a single
- table and no mapping can be specified. </para>
- <para>Invoke via:</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.WALPlayer [options] <wal inputdir> <tables> [<tableMappings>]>
-</screen>
- <para>For example:</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.WALPlayer /backuplogdir oldTable1,oldTable2 newTable1,newTable2
-</screen>
- <para> WALPlayer, by default, runs as a mapreduce job. To NOT run WALPlayer as a mapreduce job
- on your cluster, force it to run all in the local process by adding the flags
- <code>-Dmapreduce.jobtracker.address=local</code> on the command line. </para>
- </section>
- <section
- xml:id="rowcounter">
- <title>RowCounter and CellCounter</title>
- <para><link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/mapreduce/RowCounter.html">RowCounter</link>
- is a mapreduce job to count all the rows of a table. This is a good utility to use as a
- sanity check to ensure that HBase can read all the blocks of a table if there are any
- concerns of metadata inconsistency. It will run the mapreduce all in a single process but it
- will run faster if you have a MapReduce cluster in place for it to exploit.</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.RowCounter <tablename> [<column1> <column2>...]
-</screen>
- <para>Note: caching for the input Scan is configured via
- <code>hbase.client.scanner.caching</code> in the job configuration. </para>
- <para>HBase ships another diagnostic mapreduce job called <link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/mapreduce/CellCounter.html">CellCounter</link>.
- Like RowCounter, it gathers more fine-grained statistics about your table. The statistics
- gathered by RowCounter are more fine-grained and include: </para>
- <itemizedlist>
- <listitem>
- <para>Total number of rows in the table.</para>
- </listitem>
- <listitem>
- <para>Total number of CFs across all rows.</para>
- </listitem>
- <listitem>
- <para>Total qualifiers across all rows.</para>
- </listitem>
- <listitem>
- <para>Total occurrence of each CF.</para>
- </listitem>
- <listitem>
- <para>Total occurrence of each qualifier.</para>
- </listitem>
- <listitem>
- <para>Total number of versions of each qualifier.</para>
- </listitem>
- </itemizedlist>
- <para>The program allows you to limit the scope of the run. Provide a row regex or prefix to
- limit the rows to analyze. Use <code>hbase.mapreduce.scan.column.family</code> to specify
- scanning a single column family.</para>
- <screen language="bourne">$ bin/hbase org.apache.hadoop.hbase.mapreduce.CellCounter <tablename> <outputDir> [regex or prefix]</screen>
- <para>Note: just like RowCounter, caching for the input Scan is configured via
- <code>hbase.client.scanner.caching</code> in the job configuration. </para>
- </section>
- <section
- xml:id="mlockall">
- <title>mlockall</title>
- <para>It is possible to optionally pin your servers in physical memory making them less likely
- to be swapped out in oversubscribed environments by having the servers call <link
- xlink:href="http://linux.die.net/man/2/mlockall">mlockall</link> on startup. See <link
- xlink:href="https://issues.apache.org/jira/browse/HBASE-4391">HBASE-4391 Add ability to
- start RS as root and call mlockall</link> for how to build the optional library and have
- it run on startup. </para>
- </section>
- <section
- xml:id="compaction.tool">
- <title>Offline Compaction Tool</title>
- <para>See the usage for the <link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/regionserver/CompactionTool.html">Compaction
- Tool</link>. Run it like this <command>./bin/hbase
- org.apache.hadoop.hbase.regionserver.CompactionTool</command>
- </para>
- </section>
-
- <section>
- <title><command>hbase clean</command></title>
- <para>The <command>hbase clean</command> command cleans HBase data from ZooKeeper, HDFS, or
- both. It is appropriate to use for testing. Run it with no options for usage instructions.
- The <command>hbase clean</command> command was introduced in HBase 0.98.</para>
- <screen>
-$ <userinput>bin/hbase clean</userinput>
-Usage: hbase clean (--cleanZk|--cleanHdfs|--cleanAll)
-Options:
- --cleanZk cleans hbase related data from zookeeper.
- --cleanHdfs cleans hbase related data from hdfs.
- --cleanAll cleans hbase related data from both zookeeper and hdfs.
- </screen>
- </section>
- <section>
- <title><command>hbase pe</command></title>
- <para>The <command>hbase pe</command> command is a shortcut provided to run the
- <code>org.apache.hadoop.hbase.PerformanceEvaluation</code> tool, which is used for
- testing. The <command>hbase pe</command> command was introduced in HBase 0.98.4.</para>
- <para>The PerformanceEvaluation tool accepts many different options and commands. For usage
- instructions, run the command with no options.</para>
- <para>To run PerformanceEvaluation prior to HBase 0.98.4, issue the command
- <command>hbase org.apache.hadoop.hbase.PerformanceEvaluation</command>.</para>
- <para>The PerformanceEvaluation tool has received many updates in recent HBase releases,
- including support for namespaces, support for tags, cell-level ACLs and visibility labels,
- multiget support for RPC calls, increased sampling sizes, an option to randomly sleep during
- testing, and ability to "warm up" the cluster before testing starts.</para>
- </section>
- <section>
- <title><command>hbase ltt</command></title>
- <para>The <command>hbase ltt</command> command is a shortcut provided to run the
- <code>org.apache.hadoop.hbase.util.LoadTestTool</code> utility, which is used for
- testing. The <command>hbase ltt</command> command was introduced in HBase 0.98.4.</para>
- <para>You must specify either <option>-write</option> or <option>-update-read</option> as the
- first option. For general usage instructions, pass the <option>-h</option> option.</para>
- <para>To run LoadTestTool prior to HBase 0.98.4, issue the command <command>hbase
- org.apache.hadoop.hbase.util.LoadTestTool</command>.</para>
- <para>The LoadTestTool has received many updates in recent HBase releases, including support
- for namespaces, support for tags, cell-level ACLS and visibility labels, testing
- security-related features, ability to specify the number of regions per server, tests for
- multi-get RPC calls, and tests relating to replication.</para>
- </section>
- </section>
- <!-- tools -->
-
- <section
- xml:id="ops.regionmgt">
- <title>Region Management</title>
- <section
- xml:id="ops.regionmgt.majorcompact">
- <title>Major Compaction</title>
- <para>Major compactions can be requested via the HBase shell or <link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/HBaseAdmin.html#majorCompact%28java.lang.String%29">HBaseAdmin.majorCompact</link>. </para>
- <para>Note: major compactions do NOT do region merges. See <xref
- linkend="compaction" /> for more information about compactions. </para>
- </section>
- <section
- xml:id="ops.regionmgt.merge">
- <title>Merge</title>
- <para>Merge is a utility that can merge adjoining regions in the same table (see
- org.apache.hadoop.hbase.util.Merge).</para>
- <programlisting language="bourne">$ bin/hbase org.apache.hadoop.hbase.util.Merge <tablename> <region1> <region2>
-</programlisting>
- <para>If you feel you have too many regions and want to consolidate them, Merge is the utility
- you need. Merge must run be done when the cluster is down. See the <link
- xlink:href="http://ofps.oreilly.com/titles/9781449396107/performance.html">O'Reilly HBase
- Book</link> for an example of usage. </para>
- <para>You will need to pass 3 parameters to this application. The first one is the table name.
- The second one is the fully qualified name of the first region to merge, like
- "table_name,\x0A,1342956111995.7cef47f192318ba7ccc75b1bbf27a82b.". The third one is the
- fully qualified name for the second region to merge. </para>
- <para>Additionally, there is a Ruby script attached to <link
- xlink:href="https://issues.apache.org/jira/browse/HBASE-1621">HBASE-1621</link> for region
- merging. </para>
- </section>
- </section>
-
- <section
- xml:id="node.management">
- <title>Node Management</title>
- <section
- xml:id="decommission">
- <title>Node Decommission</title>
- <para>You can stop an individual RegionServer by running the following script in the HBase
- directory on the particular node:</para>
- <screen language="bourne">$ ./bin/hbase-daemon.sh stop regionserver</screen>
- <para> The RegionServer will first close all regions and then shut itself down. On shutdown,
- the RegionServer's ephemeral node in ZooKeeper will expire. The master will notice the
- RegionServer gone and will treat it as a 'crashed' server; it will reassign the nodes the
- RegionServer was carrying. </para>
- <note>
- <title>Disable the Load Balancer before Decommissioning a node</title>
- <para>If the load balancer runs while a node is shutting down, then there could be
- contention between the Load Balancer and the Master's recovery of the just decommissioned
- RegionServer. Avoid any problems by disabling the balancer first. See <xref
- linkend="lb" /> below. </para>
- </note>
- <note>
- <title xml:id="considerAsDead.sh">Kill Node Tool</title>
- <para>In hbase-2.0, in the bin directory, we added a script named
- <filename>considerAsDead.sh</filename> that can be used to kill a regionserver.
- Hardware issues could be detected by specialized monitoring tools before the
- zookeeper timeout has expired. <filename>considerAsDead.sh</filename> is a
- simple function to mark a RegionServer as dead. It deletes all the znodes
- of the server, starting the recovery process. Plug in the script into
- your monitoring/fault detection tools to initiate faster failover. Be
- careful how you use this disruptive tool. Copy the script if you need to
- make use of it in a version of hbase previous to hbase-2.0.
- </para>
- </note>
- <para> A downside to the above stop of a RegionServer is that regions could be offline for a
- good period of time. Regions are closed in order. If many regions on the server, the first
- region to close may not be back online until all regions close and after the master notices
- the RegionServer's znode gone. In Apache HBase 0.90.2, we added facility for having a node
- gradually shed its load and then shutdown itself down. Apache HBase 0.90.2 added the
- <filename>graceful_stop.sh</filename> script. Here is its usage:</para>
- <screen language="bourne">$ ./bin/graceful_stop.sh
-Usage: graceful_stop.sh [--config &conf-dir>] [--restart] [--reload] [--thrift] [--rest] &hostname>
- thrift If we should stop/start thrift before/after the hbase stop/start
- rest If we should stop/start rest before/after the hbase stop/start
- restart If we should restart after graceful stop
- reload Move offloaded regions back on to the stopped server
- debug Move offloaded regions back on to the stopped server
- hostname Hostname of server we are to stop</screen>
- <para> To decommission a loaded RegionServer, run the following: <command>$
- ./bin/graceful_stop.sh HOSTNAME</command> where <varname>HOSTNAME</varname> is the host
- carrying the RegionServer you would decommission. </para>
- <note>
- <title>On <varname>HOSTNAME</varname></title>
- <para>The <varname>HOSTNAME</varname> passed to <filename>graceful_stop.sh</filename> must
- match the hostname that hbase is using to identify RegionServers. Check the list of
- RegionServers in the master UI for how HBase is referring to servers. Its usually hostname
- but can also be FQDN. Whatever HBase is using, this is what you should pass the
- <filename>graceful_stop.sh</filename> decommission script. If you pass IPs, the script
- is not yet smart enough to make a hostname (or FQDN) of it and so it will fail when it
- checks if server is currently running; the graceful unloading of regions will not run.
- </para>
- </note>
- <para> The <filename>graceful_stop.sh</filename> script will move the regions off the
- decommissioned RegionServer one at a time to minimize region churn. It will verify the
- region deployed in the new location before it will moves the next region and so on until the
- decommissioned server is carrying zero regions. At this point, the
- <filename>graceful_stop.sh</filename> tells the RegionServer <command>stop</command>. The
- master will at this point notice the RegionServer gone but all regions will have already
- been redeployed and because the RegionServer went down cleanly, there will be no WAL logs to
- split. </para>
- <note
- xml:id="lb">
- <title>Load Balancer</title>
- <para> It is assumed that the Region Load Balancer is disabled while the
- <command>graceful_stop</command> script runs (otherwise the balancer and the
- decommission script will end up fighting over region deployments). Use the shell to
- disable the balancer:</para>
- <programlisting>hbase(main):001:0> balance_switch false
-true
-0 row(s) in 0.3590 seconds</programlisting>
- <para> This turns the balancer OFF. To reenable, do:</para>
- <programlisting>hbase(main):001:0> balance_switch true
-false
-0 row(s) in 0.3590 seconds</programlisting>
- <para>The <command>graceful_stop</command> will check the balancer and if enabled, will turn
- it off before it goes to work. If it exits prematurely because of error, it will not have
- reset the balancer. Hence, it is better to manage the balancer apart from
- <command>graceful_stop</command> reenabling it after you are done w/ graceful_stop.
- </para>
- </note>
- <section
- xml:id="draining.servers">
- <title>Decommissioning several Regions Servers concurrently</title>
- <para>If you have a large cluster, you may want to decommission more than one machine at a
- time by gracefully stopping mutiple RegionServers concurrently. To gracefully drain
- multiple regionservers at the same time, RegionServers can be put into a "draining" state.
- This is done by marking a RegionServer as a draining node by creating an entry in
- ZooKeeper under the <filename>hbase_root/draining</filename> znode. This znode has format
- <code>name,port,startcode</code> just like the regionserver entries under
- <filename>hbase_root/rs</filename> znode. </para>
- <para>Without this facility, decommissioning mulitple nodes may be non-optimal because
- regions that are being drained from one region server may be moved to other regionservers
- that are also draining. Marking RegionServers to be in the draining state prevents this
- from happening. See this <link
- xlink:href="http://inchoate-clatter.blogspot.com/2012/03/hbase-ops-automation.html">blog
- post</link> for more details.</para>
- </section>
-
- <section
- xml:id="bad.disk">
- <title>Bad or Failing Disk</title>
- <para>It is good having <xref
- linkend="dfs.datanode.failed.volumes.tolerated" /> set if you have a decent number of
- disks per machine for the case where a disk plain dies. But usually disks do the "John
- Wayne" -- i.e. take a while to go down spewing errors in <filename>dmesg</filename> -- or
- for some reason, run much slower than their companions. In this case you want to
- decommission the disk. You have two options. You can <link
- xlink:href="http://wiki.apache.org/hadoop/FAQ#I_want_to_make_a_large_cluster_smaller_by_taking_out_a_bunch_of_nodes_simultaneously._How_can_this_be_done.3F">decommission
- the datanode</link> or, less disruptive in that only the bad disks data will be
- rereplicated, can stop the datanode, unmount the bad volume (You can't umount a volume
- while the datanode is using it), and then restart the datanode (presuming you have set
- dfs.datanode.failed.volumes.tolerated > 0). The regionserver will throw some errors in its
- logs as it recalibrates where to get its data from -- it will likely roll its WAL log too
- -- but in general but for some latency spikes, it should keep on chugging. </para>
- <note>
- <title>Short Circuit Reads</title>
- <para>If you are doing short-circuit reads, you will have to move the regions off the
- regionserver before you stop the datanode; when short-circuiting reading, though chmod'd
- so regionserver cannot have access, because it already has the files open, it will be
- able to keep reading the file blocks from the bad disk even though the datanode is down.
- Move the regions back after you restart the datanode.</para>
- </note>
- </section>
- </section>
- <section
- xml:id="rolling">
- <title>Rolling Restart</title>
-
- <para>Some cluster configuration changes require either the entire cluster, or the
- RegionServers, to be restarted in order to pick up the changes. In addition, rolling
- restarts are supported for upgrading to a minor or maintenance release, and to a major
- release if at all possible. See the release notes for release you want to upgrade to, to
- find out about limitations to the ability to perform a rolling upgrade.</para>
- <para>There are multiple ways to restart your cluster nodes, depending on your situation.
- These methods are detailed below.</para>
- <section>
- <title>Using the <command>rolling-restart.sh</command> Script</title>
-
- <para>HBase ships with a script, <filename>bin/rolling-restart.sh</filename>, that allows
- you to perform rolling restarts on the entire cluster, the master only, or the
- RegionServers only. The script is provided as a template for your own script, and is not
- explicitly tested. It requires password-less SSH login to be configured and assumes that
- you have deployed using a tarball. The script requires you to set some environment
- variables before running it. Examine the script and modify it to suit your needs.</para>
- <example>
- <title><filename>rolling-restart.sh</filename> General Usage</title>
- <screen language="bourne">
-$ <userinput>./bin/rolling-restart.sh --help</userinput><![CDATA[
-Usage: rolling-restart.sh [--config <hbase-confdir>] [--rs-only] [--master-only] [--graceful] [--maxthreads xx]
- ]]></screen>
- </example>
- <variablelist>
- <varlistentry>
- <term>Rolling Restart on RegionServers Only</term>
- <listitem>
- <para>To perform a rolling restart on the RegionServers only, use the
- <code>--rs-only</code> option. This might be necessary if you need to reboot the
- individual RegionServer or if you make a configuration change that only affects
- RegionServers and not the other HBase processes.</para>
- <para>If you need to restart only a single RegionServer, or if you need to do extra
- actions during the restart, use the <filename>bin/graceful_stop.sh</filename>
- command instead. See <xref linkend="rolling.restart.manual" />.</para>
- </listitem>
- </varlistentry>
- <varlistentry>
- <term>Rolling Restart on Masters Only</term>
- <listitem>
- <para>To perform a rolling restart on the active and backup Masters, use the
- <code>--master-only</code> option. You might use this if you know that your
- configuration change only affects the Master and not the RegionServers, or if you
- need to restart the server where the active Master is running.</para>
- <para>If you are not running backup Masters, the Master is simply restarted. If you
- are running backup Masters, they are all stopped before any are restarted, to avoid
- a race condition in ZooKeeper to determine which is the new Master. First the main
- Master is restarted, then the backup Masters are restarted. Directly after restart,
- it checks for and cleans out any regions in transition before taking on its normal
- workload.</para>
- </listitem>
- </varlistentry>
- <varlistentry>
- <term>Graceful Restart</term>
- <listitem>
- <para>If you specify the <code>--graceful</code> option, RegionServers are restarted
- using the <filename>bin/graceful_stop.sh</filename> script, which moves regions off
- a RegionServer before restarting it. This is safer, but can delay the
- restart.</para>
- </listitem>
- </varlistentry>
- <varlistentry>
- <term>Limiting the Number of Threads</term>
- <listitem>
- <para>To limit the rolling restart to using only a specific number of threads, use the
- <code>--maxthreads</code> option.</para>
- </listitem>
- </varlistentry>
- </variablelist>
- </section>
- <section xml:id="rolling.restart.manual">
- <title>Manual Rolling Restart</title>
- <para>To retain more control over the process, you may wish to manually do a rolling restart
- across your cluster. This uses the <command>graceful-stop.sh</command> command <xref
- linkend="decommission" />. In this method, you can restart each RegionServer
- individually and then move its old regions back into place, retaining locality. If you
- also need to restart the Master, you need to do it separately, and restart the Master
- before restarting the RegionServers using this method. The following is an example of such
- a command. You may need to tailor it to your environment. This script does a rolling
- restart of RegionServers only. It disables the load balancer before moving the
- regions.</para>
- <screen><![CDATA[
-$ for i in `cat conf/regionservers|sort`; do ./bin/graceful_stop.sh --restart --reload --debug $i; done &> /tmp/log.txt &;
- ]]></screen>
- <para>Monitor the output of the <filename>/tmp/log.txt</filename> file to follow the
- progress of the script. </para>
- </section>
-
- <section>
- <title>Logic for Crafting Your Own Rolling Restart Script</title>
- <para>Use the following guidelines if you want to create your own rolling restart script.</para>
- <orderedlist>
- <listitem>
- <para>Extract the new release, verify its configuration, and synchronize it to all nodes
- of your cluster using <command>rsync</command>, <command>scp</command>, or another
- secure synchronization mechanism.</para></listitem>
- <listitem><para>Use the hbck utility to ensure that the cluster is consistent.</para>
- <screen>
-$ ./bin/hbck
- </screen>
- <para>Perform repairs if required. See <xref linkend="hbck" /> for details.</para>
- </listitem>
- <listitem><para>Restart the master first. You may need to modify these commands if your
- new HBase directory is different from the old one, such as for an upgrade.</para>
- <screen>
-$ ./bin/hbase-daemon.sh stop master; ./bin/hbase-daemon.sh start master
- </screen>
- </listitem>
- <listitem><para>Gracefully restart each RegionServer, using a script such as the
- following, from the Master.</para>
- <screen><![CDATA[
-$ for i in `cat conf/regionservers|sort`; do ./bin/graceful_stop.sh --restart --reload --debug $i; done &> /tmp/log.txt &
- ]]></screen>
- <para>If you are running Thrift or REST servers, pass the --thrift or --rest options.
- For other available options, run the <command>bin/graceful-stop.sh --help</command>
- command.</para>
- <para>It is important to drain HBase regions slowly when restarting multiple
- RegionServers. Otherwise, multiple regions go offline simultaneously and must be
- reassigned to other nodes, which may also go offline soon. This can negatively affect
- performance. You can inject delays into the script above, for instance, by adding a
- Shell command such as <command>sleep</command>. To wait for 5 minutes between each
- RegionServer restart, modify the above script to the following:</para>
- <screen><![CDATA[
-$ for i in `cat conf/regionservers|sort`; do ./bin/graceful_stop.sh --restart --reload --debug $i & sleep 5m; done &> /tmp/log.txt &
- ]]></screen>
- </listitem>
- <listitem><para>Restart the Master again, to clear out the dead servers list and re-enable
- the load balancer.</para></listitem>
- <listitem><para>Run the <command>hbck</command> utility again, to be sure the cluster is
- consistent.</para></listitem>
- </orderedlist>
- </section>
- </section>
- <section
- xml:id="adding.new.node">
- <title>Adding a New Node</title>
- <para>Adding a new regionserver in HBase is essentially free, you simply start it like this:
- <command>$ ./bin/hbase-daemon.sh start regionserver</command> and it will register itself
- with the master. Ideally you also started a DataNode on the same machine so that the RS can
- eventually start to have local files. If you rely on ssh to start your daemons, don't forget
- to add the new hostname in <filename>conf/regionservers</filename> on the master. </para>
- <para>At this point the region server isn't serving data because no regions have moved to it
- yet. If the balancer is enabled, it will start moving regions to the new RS. On a
- small/medium cluster this can have a very adverse effect on latency as a lot of regions will
- be offline at the same time. It is thus recommended to disable the balancer the same way
- it's done when decommissioning a node and move the regions manually (or even better, using a
- script that moves them one by one). </para>
- <para>The moved regions will all have 0% locality and won't have any blocks in cache so the
- region server will have to use the network to serve requests. Apart from resulting in higher
- latency, it may also be able to use all of your network card's capacity. For practical
- purposes, consider that a standard 1GigE NIC won't be able to read much more than
- <emphasis>100MB/s</emphasis>. In this case, or if you are in a OLAP environment and
- require having locality, then it is recommended to major compact the moved regions. </para>
-
- </section>
- </section>
- <!-- node mgt -->
-
- <section xml:id="hbase_metrics">
- <title>HBase Metrics</title>
- <para>HBase emits metrics which adhere to the <link
- xlink:href="http://hadoop.apache.org/core/docs/current/api/org/apache/hadoop/metrics/package-summary.html"
- >Hadoop metrics</link> API. Starting with HBase 0.95<footnote><para>The Metrics system was redone in
- HBase 0.96. See <link xlink:href="https://blogs.apache.org/hbase/entry/migration_to_the_new_metrics">Migration
- to the New Metrics Hotness – Metrics2</link> by Elliot Clark for detail</para></footnote>,
- HBase is configured to emit a default
- set of metrics with a default sampling period of every 10 seconds. You can use HBase
- metrics in conjunction with Ganglia. You can also filter which metrics are emitted and extend
- the metrics framework to capture custom metrics appropriate for your environment.</para>
- <section xml:id="metric_setup">
- <title>Metric Setup</title>
- <para>For HBase 0.95 and newer, HBase ships with a default metrics configuration, or
- <firstterm>sink</firstterm>. This includes a wide variety of individual metrics, and emits
- them every 10 seconds by default. To configure metrics for a given region server, edit the
- <filename>conf/hadoop-metrics2-hbase.properties</filename> file. Restart the region server
- for the changes to take effect.</para>
- <para>To change the sampling rate for the default sink, edit the line beginning with
- <literal>*.period</literal>. To filter which metrics are emitted or to extend the metrics
- framework, see <link
- xlink:href="http://hadoop.apache.org/docs/current/api/org/apache/hadoop/metrics2/package-summary.html"
- />
- </para>
- <note xml:id="rs_metrics_ganglia">
- <title>HBase Metrics and Ganglia</title>
- <para>By default, HBase emits a large number of metrics per region server. Ganglia may have
- difficulty processing all these metrics. Consider increasing the capacity of the Ganglia
- server or reducing the number of metrics emitted by HBase. See <link
- xlink:href="http://hadoop.apache.org/docs/current/api/org/apache/hadoop/metrics2/package-summary.html#filtering"
- >Metrics Filtering</link>.</para>
- </note>
- </section>
- <section>
- <title>Disabling Metrics</title>
- <para>To disable metrics for a region server, edit the
- <filename>conf/hadoop-metrics2-hbase.properties</filename> file and comment out any
- uncommented lines. Restart the region server for the changes to take effect.</para>
- </section>
-
- <section xml:id="discovering.available.metrics">
- <title>Discovering Available Metrics</title>
- <para>Rather than listing each metric which HBase emits by default, you can browse through the
- available metrics, either as a JSON output or via JMX. Different metrics are
- exposed for the Master process and each region server process.</para>
- <procedure>
- <title>Access a JSON Output of Available Metrics</title>
- <step>
- <para>After starting HBase, access the region server's web UI, at
- <literal>http://REGIONSERVER_HOSTNAME:60030</literal> by default (or port 16030 in HBase 1.0+).</para>
- </step>
- <step>
- <para>Click the <guilabel>Metrics Dump</guilabel> link near the top. The metrics for the region server are
- presented as a dump of the JMX bean in JSON format. This will dump out all metrics names and their
- values.
- To include metrics descriptions in the listing — this can be useful when you are exploring
- what is available — add a query string of
- <literal>?description=true</literal> so your URL becomes
- <literal>http://REGIONSERVER_HOSTNAME:60030/jmx?description=true</literal>.
- Not all beans and attributes have descriptions.
- </para>
- </step>
- <step>
- <para>To view metrics for the Master, connect to the Master's web UI instead (defaults to
- <literal>http://localhost:60010</literal> or port 16010 in HBase 1.0+) and click its <guilabel>Metrics
- Dump</guilabel> link.
- To include metrics descriptions in the listing — this can be useful when you are exploring
- what is available — add a query string of
- <literal>?description=true</literal> so your URL becomes
- <literal>http://REGIONSERVER_HOSTNAME:60010/jmx?description=true</literal>.
- Not all beans and attributes have descriptions.
- </para>
- </step>
- </procedure>
-
- <procedure>
- <title>Browse the JMX Output of Available Metrics</title>
- <para>You can use many different tools to view JMX content by browsing MBeans. This
- procedure uses <command>jvisualvm</command>, which is an application usually available in the JDK.
- </para>
- <step>
- <para>Start HBase, if it is not already running.</para>
- </step>
- <step>
- <para>Run the command <command>jvisualvm</command> command on a host with a GUI display.
- You can launch it from the command line or another method appropriate for your operating
- system.</para>
- </step>
- <step>
- <para>Be sure the <guilabel>VisualVM-MBeans</guilabel> plugin is installed. Browse to <menuchoice>
- <guimenu>Tools</guimenu>
- <guimenuitem>Plugins</guimenuitem>
- </menuchoice>. Click <guilabel>Installed</guilabel> and check whether the plugin is
- listed. If not, click <guilabel>Available Plugins</guilabel>, select it, and click
- <guibutton>Install</guibutton>. When finished, click
- <guibutton>Close</guibutton>.</para>
- </step>
- <step>
- <para>To view details for a given HBase process, double-click the process in the
- <guilabel>Local</guilabel> sub-tree in the left-hand panel. A detailed view opens in
- the right-hand panel. Click the <guilabel>MBeans</guilabel> tab which appears as a tab
- in the top of the right-hand panel.</para>
- </step>
- <step>
- <para>To access the HBase metrics, navigate to the appropriate sub-bean:</para>
- <itemizedlist>
- <listitem>
- <para>Master: <menuchoice>
- <guimenu>Hadoop</guimenu>
- <guisubmenu>HBase</guisubmenu>
- <guisubmenu>Master</guisubmenu>
- <guisubmenu>Server</guisubmenu>
- </menuchoice></para>
- </listitem>
- <listitem>
- <para>RegionServer: <menuchoice>
- <guimenu>Hadoop</guimenu>
- <guisubmenu>HBase</guisubmenu>
- <guisubmenu>RegionServer</guisubmenu>
- <guisubmenu>Server</guisubmenu>
- </menuchoice></para>
- </listitem>
- </itemizedlist>
- </step>
- <step>
- <para>The name of each metric and its current value is displayed in the
- <guilabel>Attributes</guilabel> tab. For a view which includes more details, including
- the description of each attribute, click the <guilabel>Metadata</guilabel> tab.</para>
- </step>
- </procedure>
- </section>
- <section>
- <title>Units of Measure for Metrics</title>
- <para>Different metrics are expressed in different units, as appropriate. Often, the unit of
- measure is in the name (as in the metric <code>shippedKBs</code>). Otherwise, use the
- following guidelines. When in doubt, you may need to examine the source for a given
- metric.</para>
- <itemizedlist>
- <listitem>
- <para>Metrics that refer to a point in time are usually expressed as a timestamp.</para>
- </listitem>
- <listitem>
- <para>Metrics that refer to an age (such as <code>ageOfLastShippedOp</code>) are usually
- expressed in milliseconds.</para>
- </listitem>
- <listitem>
- <para>Metrics that refer to memory sizes are in bytes.</para>
- </listitem>
- <listitem>
- <para>Sizes of queues (such as <code>sizeOfLogQueue</code>) are expressed as the number of
- items in the queue. Determine the size by multiplying by the block size (default is 64
- MB in HDFS).</para>
- </listitem>
- <listitem>
- <para>Metrics that refer to things like the number of a given type of operations (such as
- <code>logEditsRead</code>) are expressed as an integer.</para>
- </listitem>
- </itemizedlist>
- </section>
- <section xml:id="master_metrics">
- <title>Most Important Master Metrics</title>
- <para>Note: Counts are usually over the last metrics reporting interval.</para>
- <variablelist>
- <varlistentry>
- <term>hbase.master.numRegionServers</term>
- <listitem><para>Number of live regionservers</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.master.numDeadRegionServers</term>
- <listitem><para>Number of dead regionservers</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.master.ritCount </term>
- <listitem><para>The number of regions in transition</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.master.ritCountOverThreshold</term>
- <listitem><para>The number of regions that have been in transition longer than
- a threshold time (default: 60 seconds)</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.master.ritOldestAge</term>
- <listitem><para>The age of the longest region in transition, in milliseconds
- </para></listitem>
- </varlistentry>
- </variablelist>
- </section>
- <section xml:id="rs_metrics">
- <title>Most Important RegionServer Metrics</title>
- <para>Note: Counts are usually over the last metrics reporting interval.</para>
- <variablelist>
- <varlistentry>
- <term>hbase.regionserver.regionCount</term>
- <listitem><para>The number of regions hosted by the regionserver</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.storeFileCount</term>
- <listitem><para>The number of store files on disk currently managed by the
- regionserver</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.storeFileSize</term>
- <listitem><para>Aggregate size of the store files on disk</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.hlogFileCount</term>
- <listitem><para>The number of write ahead logs not yet archived</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.totalRequestCount</term>
- <listitem><para>The total number of requests received</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.readRequestCount</term>
- <listitem><para>The number of read requests received</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.writeRequestCount</term>
- <listitem><para>The number of write requests received</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.numOpenConnections</term>
- <listitem><para>The number of open connections at the RPC layer</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.numActiveHandler</term>
- <listitem><para>The number of RPC handlers actively servicing requests</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.numCallsInGeneralQueue</term>
- <listitem><para>The number of currently enqueued user requests</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.numCallsInReplicationQueue</term>
- <listitem><para>The number of currently enqueued operations received from
- replication</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.numCallsInPriorityQueue</term>
- <listitem><para>The number of currently enqueued priority (internal housekeeping)
- requests</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.flushQueueLength</term>
- <listitem><para>Current depth of the memstore flush queue. If increasing, we are falling
- behind with clearing memstores out to HDFS.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.updatesBlockedTime</term>
- <listitem><para>Number of milliseconds updates have been blocked so the memstore can be
- flushed</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.compactionQueueLength</term>
- <listitem><para>Current depth of the compaction request queue. If increasing, we are
- falling behind with storefile compaction.</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.blockCacheHitCount</term>
- <listitem><para>The number of block cache hits</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.blockCacheMissCount</term>
- <listitem><para>The number of block cache misses</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.blockCacheExpressHitPercent </term>
- <listitem><para>The percent of the time that requests with the cache turned on hit the
- cache</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.percentFilesLocal</term>
- <listitem><para>Percent of store file data that can be read from the local DataNode,
- 0-100</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.<op>_<measure></term>
- <listitem><para>Operation latencies, where <op> is one of Append, Delete, Mutate,
- Get, Replay, Increment; and where <measure> is one of min, max, mean, median,
- 75th_percentile, 95th_percentile, 99th_percentile</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.slow<op>Count </term>
- <listitem><para>The number of operations we thought were slow, where <op> is one
- of the list above</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.GcTimeMillis</term>
- <listitem><para>Time spent in garbage collection, in milliseconds</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.GcTimeMillisParNew</term>
- <listitem><para>Time spent in garbage collection of the young generation, in
- milliseconds</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.GcTimeMillisConcurrentMarkSweep</term>
- <listitem><para>Time spent in garbage collection of the old generation, in
- milliseconds</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.authenticationSuccesses</term>
- <listitem><para>Number of client connections where authentication succeeded</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.authenticationFailures</term>
- <listitem><para>Number of client connection authentication failures</para></listitem>
- </varlistentry>
- <varlistentry>
- <term>hbase.regionserver.mutationsWithoutWALCount </term>
- <listitem><para>Count of writes submitted with a flag indicating they should bypass the
- write ahead log</para></listitem>
- </varlistentry>
- </variablelist>
- </section>
- </section>
-
- <section
- xml:id="ops.monitoring">
- <title>HBase Monitoring</title>
- <section
- xml:id="ops.monitoring.overview">
- <title>Overview</title>
- <para>The following metrics are arguably the most important to monitor for each RegionServer
- for "macro monitoring", preferably with a system like <link
- xlink:href="http://opentsdb.net/">OpenTSDB</link>. If your cluster is having performance
- issues it's likely that you'll see something unusual with this group. </para>
- <itemizedlist>
- <title>HBase:</title>
- <listitem>
- <para>See <xref
- linkend="rs_metrics" /></para>
- </listitem>
- </itemizedlist>
-
- <itemizedlist>
- <title>OS:</title>
- <listitem>
- <para>IO Wait</para>
- </listitem>
- <listitem>
- <para>User CPU</para>
- </listitem>
- </itemizedlist>
- <itemizedlist>
- <title>Java:</title>
- <listitem>
- <para>GC</para>
- </listitem>
- </itemizedlist>
- <para> For more information on HBase metrics, see <xref
- linkend="hbase_metrics" />. </para>
- </section>
-
- <section
- xml:id="ops.slow.query">
- <title>Slow Query Log</title>
- <para>The HBase slow query log consists of parseable JSON structures describing the properties
- of those client operations (Gets, Puts, Deletes, etc.) that either took too long to run, or
- produced too much output. The thresholds for "too long to run" and "too much output" are
- configurable, as described below. The output is produced inline in the main region server
- logs so that it is easy to discover further details from context with other logged events.
- It is also prepended with identifying tags <constant>(responseTooSlow)</constant>,
- <constant>(responseTooLarge)</constant>, <constant>(operationTooSlow)</constant>, and
- <constant>(operationTooLarge)</constant> in order to enable easy filtering with grep, in
- case the user desires to see only slow queries. </para>
-
- <section>
- <title>Configuration</title>
- <para>There are two configuration knobs that can be used to adjust the thresholds for when
- queries are logged. </para>
-
- <itemizedlist>
- <listitem>
- <para><varname>hbase.ipc.warn.response.time</varname> Maximum number of milliseconds
- that a query can be run without being logged. Defaults to 10000, or 10 seconds. Can be
- set to -1 to disable logging by time. </para>
- </listitem>
- <listitem>
- <para><varname>hbase.ipc.warn.response.size</varname> Maximum byte size of response that
- a query can return without being logged. Defaults to 100 megabytes. Can be set to -1
- to disable logging by size. </para>
- </listitem>
- </itemizedlist>
- </section>
-
- <section>
- <title>Metrics</title>
- <para>The slow query log exposes to metrics to JMX.</para>
- <itemizedlist>
- <listitem>
- <para><varname>hadoop.regionserver_rpc_slowResponse</varname> a global metric reflecting
- the durations of all responses that triggered logging.</para>
- </listitem>
- <listitem>
- <para><varname>hadoop.regionserver_rpc_methodName.aboveOneSec</varname> A metric
- reflecting the durations of all responses that lasted for more than one second.</para>
- </listitem>
- </itemizedlist>
-
- </section>
-
- <section>
- <title>Output</title>
- <para>The output is tagged with operation e.g. <constant>(operationTooSlow)</constant> if
- the call was a client operation, such as a Put, Get, or Delete, which we expose detailed
- fingerprint information for. If not, it is tagged <constant>(responseTooSlow)</constant>
- and still produces parseable JSON output, but with less verbose information solely
- regarding its duration and size in the RPC itself. <constant>TooLarge</constant> is
- substituted for <constant>TooSlow</constant> if the response size triggered the logging,
- with <constant>TooLarge</constant> appearing even in the case that both size and duration
- triggered logging. </para>
- </section>
- <section>
- <title>Example</title>
- <para>
- <programlisting>2011-09-08 10:01:25,824 WARN org.apache.hadoop.ipc.HBaseServer: (operationTooSlow): {"tables":{"riley2":{"puts":[{"totalColumns":11,"families":{"actions":[{"timestamp":1315501284459,"qualifier":"0","vlen":9667580},{"timestamp":1315501284459,"qualifier":"1","vlen":10122412},{"timestamp":1315501284459,"qualifier":"2","vlen":11104617},{"timestamp":1315501284459,"qualifier":"3","vlen":13430635}]},"row":"cfcd208495d565ef66e7dff9f98764da:0"}],"families":["actions"]}},"processingtimems":956,"client":"10.47.34.63:33623","starttimems":1315501284456,"queuetimems":0,"totalPuts":1,"class":"HRegionServer","responsesize":0,"method":"multiPut"}</programlisting>
- </para>
-
- <para>Note that everything inside the "tables" structure is output produced by MultiPut's
- fingerprint, while the rest of the information is RPC-specific, such as processing time
- and client IP/port. Other client operations follow the same pattern and the same general
- structure, with necessary differences due to the nature of the individual operations. In
- the case that the call is not a client operation, that detailed fingerprint information
- will be completely absent. </para>
-
- <para>This particular example, for example, would indicate that the likely cause of slowness
- is simply a very large (on the order of 100MB) multiput, as we can tell by the "vlen," or
- value length, fields of each put in the multiPut. </para>
- </section>
- </section>
- <section>
- <title>Block Cache Monitoring</title>
- <para>Starting with HBase 0.98, the HBase Web UI includes the ability to monitor and report on
- the performance of the block cache. To view the block cache reports, click <menuchoice>
- <guimenu>Tasks</guimenu>
- <guisubmenu>Show Non-RPC Tasks</guisubmenu>
- <guimenuitem>Block Cache</guimenuitem>
- </menuchoice>. Following are a few examples of the reporting capabilities.</para>
- <figure>
- <title>Basic Info</title>
- <mediaobject>
- <imageobject>
- <imagedata fileref="bc_basic.png" width="100%"/>
- </imageobject>
- <caption>
- <para>Shows the cache implementation</para>
- </caption>
- </mediaobject>
- </figure>
- <figure>
- <title>Config</title>
- <mediaobject>
- <imageobject>
- <imagedata fileref="bc_config.png" width="100%"/>
- </imageobject>
- <caption>
- <para>Shows all cache configuration options.</para>
- </caption>
- </mediaobject>
- </figure>
- <figure>
- <title>Stats</title>
- <mediaobject>
- <imageobject>
- <imagedata fileref="bc_stats.png" width="100%"/>
- </imageobject>
- <caption>
- <para>Shows statistics about the performance of the cache.</para>
- </caption>
- </mediaobject>
- </figure>
- <figure>
- <title>L1 and L2</title>
- <mediaobject>
- <imageobject>
- <imagedata fileref="bc_l1.png" width="100%"/>
- </imageobject>
- <imageobject>
- <imagedata fileref="bc_l2_buckets.png" width="100%"/>
- </imageobject>
- <caption>
- <para>Shows information about the L1 and L2 caches.</para>
- </caption>
- </mediaobject>
- </figure>
- <para>This is not an exhaustive list of all the screens and reports available. Have a look in
- the Web UI.</para>
- </section>
-
-
-
- </section>
-
- <section
- xml:id="cluster_replication">
- <title>Cluster Replication</title>
- <note>
- <para>This information was previously available at <link
- xlink:href="http://hbase.apache.org/replication.html">Cluster Replication</link>. </para>
- </note>
- <para>HBase provides a replication mechanism to copy data between HBase
- clusters. Replication can be used as a disaster recovery solution and as a mechanism for high
- availability. You can also use replication to separate web-facing operations from back-end
- jobs such as MapReduce.</para>
-
- <para>In terms of architecture, HBase replication is master-push. This takes advantage of the
- fact that each region server has its own write-ahead log (WAL). One master cluster can
- replicate to any number of slave clusters, and each region server replicates its own stream of
- edits. For more information on the different properties of master/slave replication and other
- types of replication, see the article <link
- xlink:href="http://highscalability.com/blog/2009/8/24/how-google-serves-data-from-multiple-datacenters.html">How
- Google Serves Data From Multiple Datacenters</link>.</para>
-
- <para>Replication is asynchronous, allowing clusters to be geographically distant or to have
- some gaps in availability. This also means that data between master and slave clusters will
- not be instantly consistent. Rows inserted on the master are not immediately available or
- consistent with rows on the slave clusters. rows inserted on the master cluster won’t be
- available at the same time on the slave clusters. The goal is eventual consistency. </para>
-
- <para>The replication format used in this design is conceptually the same as the <firstterm><link
- xlink:href="http://dev.mysql.com/doc/refman/5.1/en/replication-formats.html">statement-based
- replication</link></firstterm> design used by MySQL. Instead of SQL statements, entire
- WALEdits (consisting of multiple cell inserts coming from Put and Delete operations on the
- clients) are replicated in order to maintain atomicity. </para>
-
- <para>The WALs for each region server must be kept in HDFS as long as they are needed to
- replicate data to any slave cluster. Each region server reads from the oldest log it needs to
- replicate and keeps track of the current position inside ZooKeeper to simplify failure
- recovery. That position, as well as the queue of WALs to process, may be different for every
- slave cluster.</para>
-
- <para>The clusters participating in replication can be of different sizes. The master
- cluster relies on randomization to attempt to balance the stream of replication on the slave clusters</para>
-
- <para>HBase supports master/master and cyclic replication as well as replication to multiple
- slaves.</para>
-
- <figure>
- <title>Replication Architecture Overview</title>
- <mediaobject>
- <imageobject>
- <imagedata fileref="replication_overview.png" />
- </imageobject>
- <caption>
- <para>Illustration of the replication architecture in HBase, as described in the prior
- text.</para>
- </caption>
- </mediaobject>
- </figure>
-
- <formalpara>
- <title>Enabling and Configuring Replication</title>
- <para>See the <link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/replication/package-summary.html#requirements">
- API documentation for replication</link> for information on enabling and configuring
- replication.</para>
- </formalpara>
-
- <section>
- <title>Life of a WAL Edit</title>
- <para>A single WAL edit goes through several steps in order to be replicated to a slave
- cluster.</para>
-
- <orderedlist>
- <title>When the slave responds correctly:</title>
- <listitem>
- <para>A HBase client uses a Put or Delete operation to manipulate data in HBase.</para>
- </listitem>
- <listitem>
- <para>The region server writes the request to the WAL in a way that would allow it to be
- replayed if it were not written successfully.</para>
- </listitem>
- <listitem>
- <para>If the changed cell corresponds to a column family that is scoped for replication,
- the edit is added to the queue for replication.</para>
- </listitem>
- <listitem>
- <para>In a separate thread, the edit is read from the log, as part of a batch process.
- Only the KeyValues that are eligible for replication are kept. Replicable KeyValues are
- part of a column family whose schema is scoped GLOBAL, are not part of a catalog such as
- <code>hbase:meta</code>, and did not originate from the target slave cluster, in the
- case of cyclic replication.</para>
- </listitem>
- <listitem>
- <para>The edit is tagged with the master's UUID and added to a buffer. When the buffer is
- filled, or the reader reaches the end of the file, the buffer is sent to a random region
- server on the slave cluster.</para>
- </listitem>
- <listitem>
- <para>The region server reads the edits sequentially and separates them into buffers, one
- buffer per table. After all edits are read, each buffer is flushed using <link
- xlink:href="http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/HTable.html"
- >HTable</link>, HBase's normal client. The master's UUID is preserved in the edits
- they are applied, in order to allow for cyclic replication.</para>
- </listitem>
- <listitem>
- <para>In the master, the offset for the WAL that is currently being replicated is
- registered in ZooKeeper.</para>
- </listitem>
- </orderedlist>
- <orderedlist>
- <title>When the slave does not respond:</title>
- <listitem>
- <para>The first three steps, where the edit is inserted, are identical.</para>
- </listitem>
- <listitem>
- <para>Again in a separate thread, the region server reads, filters, and edits the log
- edits in the same way as above. The slave region server does not answer the RPC
- call.</para>
- </listitem>
- <listitem>
- <para>The master sleeps and tries again a configurable number of times.</para>
- </listitem>
- <listitem>
- <para>If the slave region server is still not available, the master selects a new subset
- of region server to replicate to, and tr
<TRUNCATED>