php-doc-en/reference/mongo/connecting.xml

<?xml version="1.0" encoding="utf-8"?>
<!-- $Revision$ -->

<section xml:id="mongo.connecting" xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink">
 <title>Connecting</title>

 <para>
  Connecting to MongoDB can be as easy as <literal>new Mongo</literal>, but
  there are many additional options and configurations. The
  <function>Mongo::__construct</function> page covers all of the API options,
  but this page gives some more details and advice for practical use cases.
 </para>

 <section>
  <title>Logging In on Connection</title>
  <para>
   If MongoDB is started with the <literal>--auth</literal> or
   <literal>--keyFile</literal> options, you must log in before you can do any
   operations with the driver.  You can log in on connection by supplying the
   username and password in the connection URI:
  </para>
  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo("mongodb://${username}:${password}@localhost");

?>
]]>
  </programlisting>
  <para>
   If your connection is dropped, the driver will automatically attempt to
   reconnect and reauthenticate you.
  </para>
  <para>
   You can also authenticate on a per-database level with
   <function>MongoDB::authenticate</function>:
  </para>
  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo();
$db = $m->admin;

$db->authenticate($username, $password);

?>
]]>
  </programlisting>
  <para>
   There is a major disadvantage to this method: if the connection is dropped
   and then reconnects, the new connection will not be authenticated. If you use
   the URI format, the PHP driver will automatically authenticate the user
   whenever a new connection is made.
  </para>
  <para>
   To authenticate with a different  database, specify the database name after
   the hosts.  This example will log the user into the "blog" database:
  </para>
  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo("mongodb://${username}:${password}@localhost/blog");

?>
]]>
  </programlisting>
 </section>

 <section>
  <title>Replica Sets</title>
  <para>
   To connect to a replica set, specify one or more members of the set and use
   the <literal>replicaSet</literal> option.
  </para>
  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo("mongodb://localhost:27017", array("replicaSet" => "myReplSetName"));

?>
]]>
  </programlisting>
  <para>
   Version 1.0.9+ of the driver is required to connect to a replica set
   (earlier versions of the driver will not autodetect the master or reconnect
   correctly).
  </para>
  <para>
   The PHP driver will query the database server(s) listed to figure out who is
   master.  So long as it can connect to at least one host listed and find a
   master, the connection will succeed.  If it cannot make a connection to any
   servers listed or cannot find a master, a
   <classname>MongoConnectionException</classname> will be thrown.
  </para>
  <para>
   If the master becomes unavailable, the slaves will not promote a new master
   for a few seconds.  During that time, this connection will not be able to
   perform any database operations (connections to slaves will still be able to
   perform reads).  Thus, if you attempt to do any sort of read or write on this
   connection, it will throw an exception.
  </para>
  <para>
   Once a master is elected, attempting to perform a read or write will allow
   the driver to detect the new master. The driver will make this its primary
   database connection and continue operating normally.
  </para>
  <para>
   For more information on replica sets, see the
   <link xlink:href="&url.mongodb.replica;">core documentation</link>.
  </para>
 </section>

 <section>
  <title>Domain Socket Support</title>

  <para>
   If you are running MongoDB locally and have version 1.0.9 or better of the
   driver, you can connect to the database via file.  MongoDB automatically
   opens a socket file on startup: /tmp/mongodb-&lt;port&gt;.sock.
  </para>

  <para>
   To connect to the socket file, specify the path in your MongoDB connection
   string:
  </para>

  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo("mongodb:///tmp/mongo-27017.sock");

?>
]]>
  </programlisting>

  <para>
   If you would like to use authentication on connection (as described above)
   with a socket file, you must specify a port of 0 so that the connection
   string parser knows where the end of the connection string is.
  </para>

  <programlisting role="php">
<![CDATA[
<?php

$m = new Mongo("mongodb://username:password@/tmp/mongo-27017.sock:0/foo");

?>
]]>
  </programlisting>

 </section>

 <section>
  <title>Connection Pooling (version 1.2.0+)</title>
  <para>
   Creating connections is one of the most heavyweight things that the driver
   does. It can take hundreds of milliseconds to set up a connection correctly,
   even on a fast network.  Thus, the driver tries to minimize the number of new
   connections created by reusing connections from a pool.
  </para>
  <para>
   When a user creates a new instance of <classname>Mongo</classname>, all
   necessary connections will be taken from their pools (replica sets may
   require multple connections, one for each member of the set). When the
   <classname>Mongo</classname> instance goes out of scope, the connections will
   be returned to the pool.  When the PHP process exits, all connections in the
   pools will be closed.
  </para>
  <section>
   <title>"Why do I have so many open connections?"</title>
   <para>
    Connection pools can generate a large number of connections. This is expected
    and, using a little arithmetic, you can figure out how many connections to
    expect.  There are three factors in the total number of connections:
   </para>
   <itemizedlist>
    <listitem>
     <para>
      <literal>
       connections_per_pool
      </literal>
     </para>
     <para>
      Each connection pool will create, by default, an unlimited number of
      connections. One might assume that this is a problem: if it can create an
      unlimited number of connections, couldn't it create thousands and the
      server would run out of file descriptors? In practice, this is unlikely,
      as unused connections are returned to the pool to be used later, so future
      connections will use the same connection instead of creating a new one.
      Unless you create thousands of connections at once without letting any go
      out of scope, the number of connections open should stay at a reasonable
      number.
     </para>
     <para>
      You can see how many connections you have in a pool using the
      <function>MongoPool::info</function> function.  Add up the "in use" and
      "in pool" fields for a given server.  That is the total number of
      connections for that pool.
     </para>
    </listitem>
    <listitem>
     <para>
      <literal>
       pools_per_process
      </literal>
     </para>
     <para>
      Each MongoDB server address you're connecting to gets its own connection
      pool.  For example, if your local hostname is "example.net", connecting
      to "example.net:27017", "localhost:27017", and "/tmp/mongodb-27017.sock"
      will create three connection pools.  You can see how many connection pools
      you have open using <function>MongoPool::info</function>.
     </para>
    </listitem>
    <listitem>
     <para>
      <literal>
       processes
      </literal>
     </para>
     <para>
      Each PHP process has a separate set of pools.  PHP-FPM and Apache
      generally create between 6 and a couple dozen PHP worker children. Check
      your settings to see what the max number of PHP processes is that can be
      spawned.
     </para>
     <para>
      If you are using PHP-FPM, estimating the number of connections can be
      tricky because it will spawn more PHP-FPM workers under heavy load. To be
      on the safe side, look at the max_children parameter or add up
      spare_servers+start_servers (choose whichever number is higher).  That's
      how many PHP processes (and, thus sets of pools) you should plan for.
     </para>
    </listitem>
   </itemizedlist>
   <para>
    The three variables above can be multiplied together to give the max
    number of connections expected:
    <literal>connections_per_pool*pools_per_process*processes</literal>.  Note
    that <literal>connections_per_pool</literal> can be different for different
    pools, so <literal>connections_per_pool</literal> should be the max.
   </para>
   <para>
    For example, suppose we're getting 30 connections per pool, 10 pools per PHP
    process, and 128 PHP processes.  Then we can expect 38400 connections from
    this machine.  Thus, we should set this machine's file descriptor limit to
    be high enough to handle all of these connections or it may run out of file
    descriptors.
   </para>
   <para>
    See <classname>MongoPool</classname> for more information on connection
    pooling.
   </para>
  </section>
 </section>

 <section>
  <title>Persistent Connections</title>
  <note>
   <para>
    This section is not relevant for 1.2.0+. In 1.2.0+, connections are always
    persistent and managed automatically by the driver.  See
    <classname>MongoPool</classname> for more information on pooling.
   </para>
  </note>

  <para>
   Creating new connection to the database is very slow.  To minimize the number
   of connections that you need to make, you can use persistent connections.  A
   persistent connection is saved by PHP, so you can use the same connection for
   multiple requests.
  </para>

  <para>
   For example, this simple program connects to the database 1000 times:
  </para>

  <programlisting role="php">
<![CDATA[
<?php

for ($i=0; $i<1000; $i++) {
  $m = new Mongo();
}

?>
]]>
  </programlisting>

  <para>
   It takes approximately 18 seconds to execute.  If we change it to use a
   persistent connection:
  </para>

  <programlisting role="php">
<![CDATA[
<?php

for ($i=0; $i<1000; $i++) {
  $m = new Mongo("localhost:27017", array("persist" => "x"));
}

?>
]]>
  </programlisting>

  <para>
   ...it takes less than .02 seconds to execute, as it only makes one database
   connection.
  </para>

  <para>
   Persistent connections need an identifier string (which is "x" in the above
   example) to uniquely identify them.  For a persistent connection to be used,
   the hostname, port, persist string, and username and password (if given) must
   match an existing persistent connection.  Otherwise, a new connection will be
   created with this identifying information.
  </para>
  <para>
   Persistent connections are <emphasis>highly recommended</emphasis> and should
   always be used in production unless there is a compelling reason not to.
   Most of the reasons that they are not recommended for relational databases
   are irrelevant to MongoDB.
  </para>
 </section>

</section>