path: root/lib/kernel/doc/src/net_kernel.xml

<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE erlref SYSTEM "erlref.dtd">

<erlref>
  <header>
    <copyright>
      <year>1996</year><year>2011</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      The contents of this file are subject to the Erlang Public License,
      Version 1.1, (the "License"); you may not use this file except in
      compliance with the License. You should have received a copy of the
      Erlang Public License along with this software. If not, it can be
      retrieved online at http://www.erlang.org/.
    
      Software distributed under the License is distributed on an "AS IS"
      basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
      the License for the specific language governing rights and limitations
      under the License.
    
    </legalnotice>

    <title>net_kernel</title>
    <prepared>Claes Wikstrom</prepared>
    <docno>1</docno>
    <date>96-09-10</date>
    <rev>A</rev>
  </header>
  <module>net_kernel</module>
  <modulesummary>Erlang Networking Kernel</modulesummary>
  <description>
    <p>The net kernel is a system process, registered as
      <c>net_kernel</c>, which must be running for distributed Erlang
      to work. The purpose of this process is to implement parts of
      the BIFs <c>spawn/4</c> and <c>spawn_link/4</c>, and to provide 
      monitoring of the network.</p>
    <p>An Erlang node is started using the command line flag
      <c>-name</c> or <c>-sname</c>:</p>
    <pre>
$ <input>erl -sname foobar</input></pre>
    <p>It is also possible to call <c>net_kernel:start([foobar])</c>
      directly from the normal Erlang shell prompt:</p>
    <p></p>
    <pre>
1> <input>net_kernel:start([foobar, shortnames]).</input>
{ok,&lt;0.64.0>}
(foobar@gringotts)2></pre>
    <p>If the node is started with the command line flag <c>-sname</c>,
      the node name will be <c>foobar@Host</c>, where <c>Host</c> is
      the short name of the host (not the fully qualified domain name).
      If started with the <c>-name</c> flag, <c>Host</c> is the fully
      qualified domain name. See <c>erl(1)</c>.</p>
    <p>Normally, connections are established automatically when
      another node is referenced. This functionality can be disabled
      by setting the Kernel configuration parameter
      <c>dist_auto_connect</c> to <c>false</c>, see
      <seealso marker="kernel_app">kernel(6)</seealso>. In this case,
      connections must be established explicitly by calling
      <c>net_kernel:connect_node/1</c>.</p>
    <p>Which nodes are allowed to communicate with each other is handled
      by the magic cookie system, see
      <seealso marker="doc/reference_manual:distributed">Distributed Erlang</seealso> in the Erlang Reference Manual.</p>
  </description>
  <funcs>
    <func>
      <name name="allow" arity="1"/>
      <fsummary>Limit access to a specified set of nodes</fsummary>
      <desc>
        <p>Limits access to the specified set of nodes. Any access
          attempts made from (or to) nodes not in <c><anno>Nodes</anno></c> will be
          rejected.</p>
        <p>Returns <c>error</c> if any element in <c><anno>Nodes</anno></c> is not
          an atom.</p>
      </desc>
    </func>
    <func>
      <name name="connect_node" arity="1"/>
      <fsummary>Establish a connection to a node</fsummary>
      <desc>
        <p>Establishes a connection to <c><anno>Node</anno></c>. Returns <c>true</c>
          if successful, <c>false</c> if not, and <c>ignored</c> if
          the local node is not alive.</p>
      </desc>
    </func>
    <func>
      <name name="monitor_nodes" arity="1"/>
      <name name="monitor_nodes" arity="2"/>
      <fsummary>Subscribe to node status change messages</fsummary>
      <desc>
        <p>The calling process subscribes or unsubscribes to node
          status change messages. A <c>nodeup</c> message is delivered
          to all subscribing process when a new node is connected, and
          a <c>nodedown</c> message is delivered when a node is
          disconnected.</p>
        <p>If <c><anno>Flag</anno></c> is <c>true</c>, a new subscription is started.
          If <c><anno>Flag</anno></c> is <c>false</c>, all previous subscriptions --
          started with the same <c><anno>Options</anno></c> -- are stopped. Two
          option lists are considered the same if they contain the same
          set of options.</p>
        <p>As of <c>kernel</c> version 2.11.4, and <c>erts</c> version
          5.5.4, the following is guaranteed:</p>
        <list type="bulleted">
          <item><c>nodeup</c> messages will be delivered before delivery
           of any message from the remote node passed through the
           newly established connection.</item>
          <item><c>nodedown</c> messages will not be delivered until all
           messages from the remote node that have been passed
           through the connection have been delivered.</item>
        </list>
        <p>Note, that this is <em>not</em> guaranteed for <c>kernel</c>
          versions before 2.11.4.</p>
        <p>As of <c>kernel</c> version 2.11.4 subscriptions can also be
          made before the <c>net_kernel</c> server has been started,
          i.e., <c>net_kernel:monitor_nodes/[1,2]</c> does not return
          <c>ignored</c>.</p>
        <p>As of <c>kernel</c> version 2.13, and <c>erts</c> version
          5.7, the following is guaranteed:</p>
        <list type="bulleted">
          <item><c>nodeup</c> messages will be delivered after the
	  corresponding node appears in results from
	  <c>erlang:nodes/X</c>.</item>
          <item><c>nodedown</c> messages will be delivered after the
	  corresponding node has disappeared in results from
	  <c>erlang:nodes/X</c>.</item>
        </list>
        <p>Note, that this is <em>not</em> guaranteed for <c>kernel</c>
          versions before 2.13.</p>
        <p>The format of the node status change messages depends on
          <c><anno>Options</anno></c>. If <c><anno>Options</anno></c> is [], which is the default,
          the format is:</p>
        <code type="none">
{nodeup, Node} | {nodedown, Node}
  Node = node()</code>
        <p>If <c><anno>Options</anno> /= []</c>, the format is:</p>
        <code type="none">
{nodeup, Node, InfoList} | {nodedown, Node, InfoList}
  Node = node()
  InfoList = [{Tag, Val}]</code>
        <p><c>InfoList</c> is a list of tuples. Its contents depends on
          <c><anno>Options</anno></c>, see below.</p>
        <p>Also, when <c>OptionList == []</c> only visible nodes, that
          is, nodes that appear in the result of
          <seealso marker="erts:erlang#nodes/0">nodes/0</seealso>, are
          monitored.</p>
        <p><c><anno>Option</anno></c> can be any of the following:</p>
        <taglist>
          <tag><c>{node_type, NodeType}</c></tag>
          <item>
            <p>Currently valid values for <c>NodeType</c> are:</p>
            <taglist>
              <tag><c>visible</c></tag>
              <item>Subscribe to node status change messages for visible
               nodes only. The tuple <c>{node_type, visible}</c> is
               included in <c>InfoList</c>.</item>
              <tag><c>hidden</c></tag>
              <item>Subscribe to node status change messages for hidden
               nodes only. The tuple <c>{node_type, hidden}</c> is
               included in <c>InfoList</c>.</item>
              <tag><c>all</c></tag>
              <item>Subscribe to node status change messages for both
               visible and hidden nodes. The tuple
              <c>{node_type, visible | hidden}</c> is included in
              <c>InfoList</c>.</item>
            </taglist>
          </item>
          <tag><c>nodedown_reason</c></tag>
          <item>
            <p>The tuple <c>{nodedown_reason, Reason}</c> is included in
              <c>InfoList</c> in <c>nodedown</c> messages. <c>Reason</c>
              can be:</p>
            <taglist>
              <tag><c>connection_setup_failed</c></tag>
              <item>The connection setup failed (after <c>nodeup</c>
               messages had been sent).</item>
              <tag><c>no_network</c></tag>
              <item>No network available.</item>
              <tag><c>net_kernel_terminated</c></tag>
              <item>The <c>net_kernel</c> process terminated.</item>
              <tag><c>shutdown</c></tag>
              <item>Unspecified connection shutdown.</item>
              <tag><c>connection_closed</c></tag>
              <item>The connection was closed.</item>
              <tag><c>disconnect</c></tag>
              <item>The connection was disconnected (forced from the
               current node).</item>
              <tag><c>net_tick_timeout</c></tag>
              <item>Net tick timeout.</item>
              <tag><c>send_net_tick_failed</c></tag>
              <item>Failed to send net tick over the connection.</item>
              <tag><c>get_status_failed</c></tag>
              <item>Status information retrieval from the <c>Port</c>
               holding the connection failed.</item>
            </taglist>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="get_net_ticktime" arity="0"/>
      <fsummary>Get <c>net_ticktime</c></fsummary>
      <desc>
        <p>Gets <c>net_ticktime</c> (see
          <seealso marker="kernel_app">kernel(6)</seealso>).</p>
        <p>Currently defined return values (<c><anno>Res</anno></c>):</p>
        <taglist>
          <tag><c><anno>NetTicktime</anno></c></tag>
          <item>
            <p><c>net_ticktime</c> is <c><anno>NetTicktime</anno></c> seconds.</p>
          </item>
          <tag><c>{ongoing_change_to, <anno>NetTicktime</anno>}</c></tag>
          <item>
            <p><c>net_kernel</c> is currently changing
              <c>net_ticktime</c> to <c><anno>NetTicktime</anno></c> seconds.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name name="set_net_ticktime" arity="1"/>
      <name name="set_net_ticktime" arity="2"/>
      <fsummary>Set <c>net_ticktime</c></fsummary>
      <desc>
        <p>Sets <c>net_ticktime</c> (see
          <seealso marker="kernel_app">kernel(6)</seealso>) to
          <c><anno>NetTicktime</anno></c> seconds. <c><anno>TransitionPeriod</anno></c> defaults
          to 60.</p>
        <p>Some definitions:</p>
        <p></p>
        <taglist>
          <tag>The minimum transition traffic interval (<c>MTTI</c>)</tag>
          <item>
            <p><c>minimum(<anno>NetTicktime</anno>, PreviousNetTicktime)*1000 div 4</c> milliseconds.</p>
          </item>
          <tag>The transition period</tag>
          <item>
            <p>The time of the least number of consecutive <c>MTTI</c>s
              to cover <c><anno>TransitionPeriod</anno></c> seconds following
              the call to <c>set_net_ticktime/2</c> (i.e.
              ((<c><anno>TransitionPeriod</anno>*1000 - 1) div MTTI + 1)*MTTI</c>
              milliseconds).</p>
          </item>
        </taglist>
        <p>If <c><![CDATA[<anno>NetTicktime</anno> < PreviousNetTicktime]]></c>, the actual
          <c>net_ticktime</c> change will be done at the end of
          the transition period; otherwise, at the beginning. During
          the transition period, <c>net_kernel</c> will ensure that
          there will be outgoing traffic on all connections at least
          every <c>MTTI</c> millisecond.</p>
        <note>
          <p>The <c>net_ticktime</c> changes have to be initiated on all
            nodes in the network (with the same <c><anno>NetTicktime</anno></c>)
            before the end of any transition period on any node;
            otherwise, connections may erroneously be disconnected.</p>
        </note>
        <p>Returns one of the following:</p>
        <taglist>
          <tag><c>unchanged</c></tag>
          <item>
            <p><c>net_ticktime</c> already had the value of
              <c><anno>NetTicktime</anno></c> and was left unchanged.</p>
          </item>
          <tag><c>change_initiated</c></tag>
          <item>
            <p><c>net_kernel</c> has initiated the change of
              <c>net_ticktime</c> to <c><anno>NetTicktime</anno></c> seconds.</p>
          </item>
          <tag><c>{ongoing_change_to, <anno>NewNetTicktime</anno>}</c></tag>
          <item>
            <p>The request was <em>ignored</em>; because,
              <c>net_kernel</c> was busy changing <c>net_ticktime</c> to
              <c><anno>NewNetTicktime</anno></c> seconds.</p>
          </item>
        </taglist>
      </desc>
    </func>
    <func>
      <name>start([Name]) -> {ok, pid()} | {error, Reason}</name>
      <name>start([Name, NameType]) -> {ok, pid()} | {error, Reason}</name>
      <name>start([Name, NameType, Ticktime]) -> {ok, pid()} | {error, Reason}</name>
      <fsummary>Turn an Erlang runtime system into a distributed node</fsummary>
      <type>
        <v>Name = atom()</v>
        <v>NameType = shortnames | longnames</v>
        <v>Reason = {already_started, pid()} | term()</v>
      </type>
      <desc>
        <p>Note that the argument is a list with exactly one, two or
          three arguments. <c>NameType</c> defaults to <c>longnames</c>
          and <c>Ticktime</c> to 15000.</p>
        <p>Turns a non-distributed node into a distributed node by
          starting <c>net_kernel</c> and other necessary processes.</p>
      </desc>
    </func>
    <func>
      <name name="stop" arity="0"/>
      <fsummary>Turn a node into a non-distributed Erlang runtime system</fsummary>
      <desc>
        <p>Turns a distributed node into a non-distributed node. For
          other nodes in the network, this is the same as the node
          going down. Only possible when the net kernel was started
          using <c>start/1</c>, otherwise returns
          <c>{error, not_allowed}</c>. Returns <c>{error, not_found}</c>
          if the local node is not alive.</p>
      </desc>
    </func>
  </funcs>
</erlref>