path: root/doc/src/sgml/example-basic.sgml

<!-- doc/src/sgml/example-basic.sgml -->
<sect1 id="example-basic">
 <title>Basic Configuration Example</title>

 <sect2 id="example-configs-begin">
  <title>Let's Begin!</title>
  <para>
   First, we must learn how to install and configure <productname>Pgpool-II</productname> and database nodes before using replication.
  </para>

  <sect3 id="example-configs-begin-installing">
   <title>Installing <productname>Pgpool-II</productname></title>
   <para>
    Installing <productname>Pgpool-II</productname> is very easy.
    In the directory which you have extracted the source tar ball,
    execute the following commands.
    <programlisting>
     $ ./configure
     $ make
     $ make install
    </programlisting>
    <command>configure</command> script collects your system information
    and use it for the compilation procedure. You can pass command
    line arguments to <command>configure</command> script to change the default behavior,
    such as the installation directory. <productname>Pgpool-II</productname>
    will be installed to <literal>/usr/local</literal> directory by default.
   </para>
   <para>
    <command>make</command> command compiles the source code, and
    <command>make install</command> will install the executables.
    You must have write permission on the installation directory.
    In this tutorial, we will install <productname>Pgpool-II
    </productname> in the default <literal>/usr/local</literal> directory.
   </para>
   <note>
    <para>
     <productname>Pgpool-II</productname> requires <literal>libpq</literal>
     library in <productname>PostgreSQL</productname> 7.4 or later (version 3 protocol).
    </para>
   </note>
   <para>
    If the <command>configure</command> script displays the following error message, the
    <literal>libpq</literal> library may not be installed, or it is not of version 3
    <programlisting>
     configure: error: libpq is not installed or libpq is old
    </programlisting>
    If the library is version 3, but the above message is still displayed, your
    <literal>libpq</literal> library is probably not recognized by the <command>
     configure</command> script.
    The <command>configure</command> script searches for <literal>libpq</literal>
    library under <literal>/usr/local/pgsql</literal>. If you have installed the
    <productname>PostgreSQL</productname> in a directory other than <literal>/usr/local/pgsql</literal>, use
    <literal>--with-pgsql</literal>, or <literal>--with-pgsql-includedir</literal>
    and <literal>--with-pgsql-libdir</literal> command line options when you
    execute <command>configure</command>.
   </para>
  </sect3>

  <sect3 id="example-configs-begin-config-files">
   <title>Configuration Files</title>
   <para>
    <productname>Pgpool-II</productname> configuration parameters are saved in the
    <literal>pgpool.conf</literal> file. The file is in <literal>"parameter = value"
    </literal> per line format. When you install <productname>Pgpool-II</productname>,
    <literal>pgpool.conf.sample</literal> is automatically created.
    We recommend copying and renaming it to <literal>pgpool.conf</literal>, and edit
    it as you like.
    <programlisting>
     $ cp /usr/local/etc/pgpool.conf.sample /usr/local/etc/pgpool.conf
    </programlisting>
    <productname>Pgpool-II</productname> only accepts connections from the localhost
    using port 9999 by the default. If you wish to receive connections from other hosts,
    set <xref linkend="guc-listen-addresses"> to <literal>'*'</literal>.
     <programlisting>
      listen_addresses = 'localhost'
      port = 9999
     </programlisting>
     We will use the default parameters in this tutorial.
   </para>
  </sect3>

  <sect3 id="example-configs-begin-config-pcp">
   <title>Configuring <acronym>PCP</acronym> Commands</title>
   <para>
    <productname>Pgpool-II</productname> has an interface for administrative
    purpose to retrieve information on database nodes, shutdown
    <productname>Pgpool-II</productname>, etc. via network. To use
    <acronym>PCP</acronym> commands, user authentication is required.
    This authentication is different from <productname>PostgreSQL</productname>'s user authentication.
    A user name and password need to be defined in the <literal>pcp.conf</literal>
    file. In the file, a user name and password are listed as a pair on each line,
    and they are separated by a colon (:). Passwords are encrypted in
    <literal>md5</literal> hash format.

    <programlisting>
     postgres:e8a48653851e28c69d0506508fb27fc5
    </programlisting>

    When you install <productname>Pgpool-II</productname>, <literal>pcp.conf.sample
    </literal> is automatically created. We recommend copying and renaming it
    to <literal>pcp.conf</literal>, and edit it.
    <programlisting>
     $ cp /usr/local/etc/pcp.conf.sample /usr/local/etc/pcp.conf
    </programlisting>
    To encrypt your password into md5 hash format, use the <command>pg_md5</command>
    command, which is installed as one of <productname>Pgpool-II</productname>'s
    executables. <command>pg_md5</command> takes text as a command line argument,
    and displays its md5-hashed text.
    For example, give <literal>"postgres"</literal> as the command line argument,
    and <command>pg_md5</command> displays md5-hashed text on its standard output.
    <programlisting>
     $ /usr/local/bin/pg_md5 postgres
     e8a48653851e28c69d0506508fb27fc5
    </programlisting>
    PCP commands are executed via network, so the port number must be configured
    with <xref linkend="guc-pcp-port"> parameter in <literal>pgpool.conf</literal> file.
     We will use the default 9898 for <xref linkend="guc-pcp-port"> in this tutorial.
      <programlisting>
       pcp_port = 9898
      </programlisting>
   </para>
  </sect3>

  <sect3 id="example-configs-prep-db-nodes">
   <title>Preparing Database Nodes</title>
   <para>
    Now, we need to set up backend <productname>PostgreSQL</productname> servers for <productname>Pgpool-II
    </productname>. These servers can be placed within the same host as
    <productname>Pgpool-II</productname>, or on separate machines. If you decide
    to place the servers on the same host, different port numbers must be assigned
    for each server. If the servers are placed on separate machines,
    they must be configured properly so that they can accept network
    connections from <productname>Pgpool-II</productname>.
    In this example, we create 3 <productname>PostgreSQL</productname> servers and
    specify the PostgreSQL information in the following parameters.

    <programlisting>
     backend_hostname0 = 'localhost'
     backend_port0 = 5432
     backend_weight0 = 1
     backend_hostname1 = 'localhost'
     backend_port1 = 5433
     backend_weight1 = 1
     backend_hostname2 = 'localhost'
     backend_port2 = 5434
     backend_weight2 = 1
    </programlisting>

    For <xref linkend="guc-backend-hostname">, <xref linkend="guc-backend-port">,
      <xref linkend="guc-backend-weight">, set the node's hostname, port number,
       and ratio for load balancing. At the end of each parameter string,
       node ID must be specified by adding positive integers starting with 0 (i.e. 0, 1, 2..).
   </para>
   <note>
    <para>
     <xref linkend="guc-backend-weight"> parameters for all nodes are
      set to 1, meaning that SELECT queries are equally distributed among
      three servers.
    </para>
   </note>
  </sect3>

  <sect3 id="example-configs-start-stop-pgpool">
   <title>Starting/Stopping <productname>Pgpool-II</productname></title>
   <para>
    To fire up <productname>Pgpool-II</productname>, execute the following
    command on a terminal.

    <programlisting>
     $ pgpool
    </programlisting>

    The above command, however, prints no log messages because <productname>
     Pgpool-II</productname> detaches the terminal. If you want to show
    <productname>Pgpool-II</productname> log messages, you pass <literal>-n</literal>
    option to <command>pgpool</command> command so <productname>Pgpool-II</productname>
    is executed as non-daemon process, and the terminal will not be detached.
    <programlisting>
     $ pgpool -n &
    </programlisting>

    The log messages are printed on the terminal, so it is recommended to use the following options.
    <programlisting>
     $ pgpool -n -d > /tmp/pgpool.log 2>&1 &
    </programlisting>

    The <literal>-d</literal> option enables debug messages to be generated.
    The above command keeps appending log messages to <literal>/tmp/pgpool.log
    </literal>. If you need to rotate log files, pass the logs to a external
    command which has log rotation function.
    For example, you can use <ulink url="https://httpd.apache.org/docs/2.4/programs/rotatelogs.html">
     <command>rotatelogs</command></ulink> from Apache2:
    <programlisting>
     $ pgpool -n 2>&1 | /usr/local/apache2/bin/rotatelogs \
     -l -f /var/log/pgpool/pgpool.log.%A 86400 &
    </programlisting>

    This will generate a log file named <literal>"pgpool.log.Thursday"</literal>
    then rotate it 00:00 at midnight. Rotatelogs adds logs to a file if it already
    exists. To delete old log files before rotation, you could use cron:
    <programlisting>
     55 23 * * * /usr/bin/find /var/log/pgpool -type f -mtime +5 -exec /bin/rm -f '{}' \;
    </programlisting>

    Please note that rotatelogs may exist as <literal>/usr/sbin/rotatelogs2</literal>
    in some distributions. <literal>-f</literal> option generates a log file as soon as
    <command>rotatelogs</command> starts and is available in apache2 2.2.9 or greater.
    Also <ulink url="http://www.cronolog.org/">cronolog</ulink> can be used.
    <programlisting>
     $ pgpool -n 2>&1 | /usr/sbin/cronolog \
     --hardlink=/var/log/pgsql/pgpool.log \
     '/var/log/pgsql/%Y-%m-%d-pgpool.log' &
    </programlisting>

    To stop <productname>Pgpool-II</productname>  execute the following command.
    <programlisting>
     $ pgpool stop
    </programlisting>

    If any client is still connected, <productname>Pgpool-II</productname>
    waits for it to disconnect, and then terminates itself. Run the following
    command instead if you want to shutdown <productname>Pgpool-II</productname>
    forcibly.
    <programlisting>
     $ pgpool -m fast stop
    </programlisting>

   </para>
  </sect3>
 </sect2>

 <sect2 id="example-configs-replication">
  <title>Your First Replication</title>
  <para>
   Replication (see <xref linkend="guc-replication-mode">) enables
    the same data to be copied to multiple database nodes.
    In this section, we'll use three database nodes, which we have already set
    up in <xref linkend="example-configs-begin">, and takes you step by step to
     create a database replication system.
     Sample data to be replicated will be generated by the
     <ulink url="https://www.postgresql.org/docs/current/static/pgbench.html">
      <command>pgbench</command></ulink> benchmark program.
  </para>

  <sect3 id="example-configs-config-replication">
   <title>Configuring Replication</title>
   <para>
    To enable the database replication function, set
    <xref linkend="guc-replication-mode"> to on in <literal>pgpool.conf</literal> file.
     <programlisting>
      replication_mode = true
     </programlisting>
     When <xref linkend="guc-replication-mode"> is on, <productname>Pgpool-II</productname>
      will send a copy of a received query to all the database nodes.
      In addition, when <xref linkend="guc-load-balance-mode"> is set to true,
       <productname>Pgpool-II</productname> will distribute <acronym>SELECT</acronym> queries
       among the database nodes.
       <programlisting>
	load_balance_mode = true
       </programlisting>
       In this section, we will enable both <xref linkend="guc-replication-mode">
	and <xref linkend="guc-load-balance-mode">.
   </para>
  </sect3>

  <sect3 id="example-configs-checking-replication">
   <title>Checking Replication</title>
   <para>
    To reflect the above changes in <literal>pgpool.conf</literal>,
    <productname>Pgpool-II</productname> must be restarted.
    Please refer to "Starting/Stopping <productname>Pgpool-II</productname>"
    <xref linkend="example-configs-start-stop-pgpool">.
     After configuring <literal>pgpool.conf</literal> and restarting the
     <productname>Pgpool-II</productname>, let's try the actual replication
     and see if everything is working.
     First, we need to create a database to be replicated. We will name it
     <literal>"bench_replication"</literal>. This database needs to be created
     on all the nodes. Use the
     <ulink url="https://www.postgresql.org/docs/current/static/app-createdb.html">
      <command>createdb</command></ulink> commands through
     <productname>Pgpool-II</productname>, and the database will be created
     on all the nodes.
     <programlisting>
      $ createdb -p 9999 bench_replication
     </programlisting>
     Then, we'll execute <ulink url="https://www.postgresql.org/docs/current/static/pgbench.html">
      <command>pgbench</command></ulink> with <literal>-i</literal> option.
     <literal>-i</literal> option initializes the database with pre-defined tables and data.
     <programlisting>
      $ pgbench -i -p 9999 bench_replication
     </programlisting>
     The following table is the summary of tables and data, which will be created by
     <ulink url="https://www.postgresql.org/docs/current/static/pgbench.html">
      <command>pgbench -i</command></ulink>. If, on all the nodes, the listed tables and
     data are created, replication is working correctly.
   </para>

   <table id="example-configs-checking-replication-table">
    <title>data summary</title>
    <tgroup cols="2">
     <thead>
      <row>
       <entry>Table Name</entry>
       <entry>Number of Rows</entry>
      </row>
     </thead>

     <tbody>
      <row>
       <entry>pgbench_branches</entry>
       <entry>1</entry>
      </row>

      <row>
       <entry>pgbench_tellers</entry>
       <entry>10</entry>
      </row>

      <row>
       <entry>pgbench_accounts</entry>
       <entry>100000</entry>
      </row>

      <row>
       <entry>pgbench_history</entry>
       <entry>0</entry>
      </row>

     </tbody>
    </tgroup>
   </table>

   <para>
    Let's use a simple shell script to check the above on all the nodes.
    The following script will display the number of rows in pgbench_branches,
    pgbench_tellers, pgbench_accounts, and pgbench_history tables on all the nodes (5432, 5433, 5434).
    <programlisting>
     $ for port in 5432 5433 5434; do
     >     echo $port
     >     for table_name in pgbench_branches pgbench_tellers pgbench_accounts pgbench_history; do
     >         echo $table_name
     >         psql -c "SELECT count(*) FROM $table_name" -p $port bench_replication
     >     done
     > done
    </programlisting>

   </para>
  </sect3>
 </sect2>

</sect1>