<?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>supervisor</title>
<prepared></prepared>
<docno></docno>
<date></date>
<rev></rev>
</header>
<module>supervisor</module>
<modulesummary>Generic Supervisor Behaviour</modulesummary>
<description>
<p>A behaviour module for implementing a supervisor, a process which
supervises other processes called child processes. A child
process can either be another supervisor or a worker process.
Worker processes are normally implemented using one of
the <c>gen_event</c>, <c>gen_fsm</c>, or <c>gen_server</c>
behaviours. A supervisor implemented using this module will have
a standard set of interface functions and include functionality
for tracing and error reporting. Supervisors are used to build an
hierarchical process structure called a supervision tree, a
nice way to structure a fault tolerant application. Refer to
<em>OTP Design Principles</em> for more information.</p>
<p>A supervisor assumes the definition of which child processes to
supervise to be located in a callback module exporting a
pre-defined set of functions.</p>
<p>Unless otherwise stated, all functions in this module will fail
if the specified supervisor does not exist or if bad arguments
are given.</p>
</description>
<section>
<title>Supervision Principles</title>
<p>The supervisor is responsible for starting, stopping and
monitoring its child processes. The basic idea of a supervisor is
that it should keep its child processes alive by restarting them
when necessary.</p>
<p>The children of a supervisor is defined as a list of <em>child specifications</em>. When the supervisor is started, the child
processes are started in order from left to right according to
this list. When the supervisor terminates, it first terminates
its child processes in reversed start order, from right to left.</p>
<p>A supervisor can have one of the following <em>restart strategies</em>:</p>
<list type="bulleted">
<item>
<p><c>one_for_one</c> - if one child process terminates and
should be restarted, only that child process is affected.</p>
</item>
<item>
<p><c>one_for_all</c> - if one child process terminates and
should be restarted, all other child processes are terminated
and then all child processes are restarted.</p>
</item>
<item>
<p><c>rest_for_one</c> - if one child process terminates and
should be restarted, the 'rest' of the child processes --
i.e. the child processes after the terminated child process
in the start order -- are terminated. Then the terminated
child process and all child processes after it are restarted.</p>
</item>
<item>
<p><c>simple_one_for_one</c> - a simplified <c>one_for_one</c>
supervisor, where all child processes are dynamically added
instances of the same process type, i.e. running the same
code.</p>
<p>The functions <c>delete_child/2</c>
and <c>restart_child/2</c> are invalid for
<c>simple_one_for_one</c> supervisors and will return
<c>{error,simple_one_for_one}</c> if the specified supervisor
uses this restart strategy.</p>
<p>The function <c>terminate_child/2</c> can be used for
children under <c>simple_one_for_one</c> supervisors by
giving the child's <c>pid()</c> as the second argument. If
instead the child specification identifier is used,
<c>terminate_child/2</c> will return
<c>{error,simple_one_for_one}</c>.</p>
</item>
</list>
<p>To prevent a supervisor from getting into an infinite loop of
child process terminations and restarts, a <em>maximum restart frequency</em> is defined using two integer values <c>MaxR</c>
and <c>MaxT</c>. If more than <c>MaxR</c> restarts occur within
<c>MaxT</c> seconds, the supervisor terminates all child
processes and then itself.
</p>
<marker id="child_spec"/>
<p>This is the type definition of a child specification:</p>
<pre>
child_spec() = {Id,StartFunc,Restart,Shutdown,Type,Modules}
Id = term()
StartFunc = {M,F,A}
M = F = atom()
A = [term()]
Restart = permanent | transient | temporary
Shutdown = brutal_kill | int()>=0 | infinity
Type = worker | supervisor
Modules = [Module] | dynamic
Module = atom()</pre>
<list type="bulleted">
<item>
<p><c>Id</c> is a name that is used to identify the child
specification internally by the supervisor.</p>
</item>
<item>
<p><c>StartFunc</c> defines the function call used to start
the child process. It should be a module-function-arguments
tuple <c>{M,F,A}</c> used as <c>apply(M,F,A)</c>.</p>
<p> <br></br>
</p>
<p>The start function <em>must create and link to</em> the child
process, and should return <c>{ok,Child}</c> or
<c>{ok,Child,Info}</c> where <c>Child</c> is the pid of
the child process and <c>Info</c> an arbitrary term which is
ignored by the supervisor.</p>
<p> <br></br>
</p>
<p>The start function can also return <c>ignore</c> if the child
process for some reason cannot be started, in which case
the child specification will be kept by the supervisor but
the non-existing child process will be ignored.</p>
<p> <br></br>
</p>
<p>If something goes wrong, the function may also return an
error tuple <c>{error,Error}</c>.</p>
<p> <br></br>
</p>
<p>Note that the <c>start_link</c> functions of the different
behaviour modules fulfill the above requirements.</p>
</item>
<item>
<p><c>Restart</c> defines when a terminated child process
should be restarted. A <c>permanent</c> child process should
always be restarted, a <c>temporary</c> child process should
never be restarted (even when the supervisor's restart strategy
is <c>rest_for_one</c> or <c>one_for_all</c> and a sibling's
death causes the temporary process to be terminated) and a
<c>transient</c> child process should be restarted only if
it terminates abnormally, i.e. with another exit reason
than <c>normal</c>.</p>
</item>
<item>
<p><c>Shutdown</c> defines how a child process should be
terminated. <c>brutal_kill</c> means the child process will
be unconditionally terminated using <c>exit(Child,kill)</c>.
An integer timeout value means that the supervisor will tell
the child process to terminate by calling
<c>exit(Child,shutdown)</c> and then wait for an exit signal
with reason <c>shutdown</c> back from the child process. If
no exit signal is received within the specified number of milliseconds,
the child process is unconditionally terminated using
<c>exit(Child,kill)</c>.</p>
<p>If the child process is another supervisor, <c>Shutdown</c>
should be set to <c>infinity</c> to give the subtree ample
time to shutdown. It is also allowed to set it to <c>infinity</c>,
if the child process is a worker.</p>
<warning>
<p>Be careful by setting the <c>Shutdown</c> strategy to
<c>infinity</c> when the child process is a worker. Because, in this
situation, the termination of the supervision tree depends on the
child process, it must be implemented in a safe way and its cleanup
procedure must always return.</p>
</warning>
<p><em>Important note on simple-one-for-one supervisors:</em>
The dynamically created child processes of a
simple-one-for-one supervisor are not explicitly killed,
regardless of shutdown strategy, but are expected to terminate
when the supervisor does (that is, when an exit signal from
the parent process is received).</p>
<p>Note that all child processes implemented using the standard
OTP behavior modules automatically adhere to the shutdown
protocol.</p>
</item>
<item>
<p><c>Type</c> specifies if the child process is a supervisor or
a worker.</p>
</item>
<item>
<p><c>Modules</c> is used by the release handler during code
replacement to determine which processes are using a certain
module. As a rule of thumb <c>Modules</c> should be a list
with one element <c>[Module]</c>, where <c>Module</c> is
the callback module, if the child process is a supervisor,
gen_server or gen_fsm. If the child process is an event
manager (gen_event) with a dynamic set of callback modules,
<c>Modules</c> should be <c>dynamic</c>. See <em>OTP Design Principles</em> for more information about release handling.</p>
</item>
<item>
<p>Internally, the supervisor also keeps track of the pid
<c>Child</c> of the child process, or <c>undefined</c> if no
pid exists.</p>
</item>
</list>
</section>
<datatypes>
<datatype>
<name name="child"/>
</datatype>
<datatype>
<name name="child_id"/>
<desc><p>Not a <c>pid()</c>.</p></desc>
</datatype>
<datatype>
<name name="child_spec"/>
</datatype>
<datatype>
<name name="mfargs"/>
<desc><p><c>A</c> (the argument list) has the value
<c>undefined</c> if <c>Restart</c> is <c>temporary</c>.</p>
</desc>
</datatype>
<datatype>
<name name="modules"/>
</datatype>
<datatype>
<name name="restart"/>
</datatype>
<datatype>
<name name="shutdown"/>
</datatype>
<datatype>
<name name="strategy"/>
</datatype>
<datatype>
<name name="sup_ref"/>
</datatype>
<datatype>
<name name="worker"/>
</datatype>
</datatypes>
<funcs>
<func>
<name name="start_link" arity="2"/>
<name name="start_link" arity="3"/>
<fsummary>Create a supervisor process.</fsummary>
<type name="startlink_ret"/>
<type name="startlink_err"/>
<type name="sup_name"/>
<desc>
<p>Creates a supervisor process as part of a supervision tree.
The function will, among other things, ensure that
the supervisor is linked to the calling process (its
supervisor).</p>
<p>The created supervisor process calls <c><anno>Module</anno>:init/1</c> to
find out about restart strategy, maximum restart frequency
and child processes. To ensure a synchronized start-up
procedure, <c>start_link/2,3</c> does not return until
<c><anno>Module</anno>:init/1</c> has returned and all child processes
have been started.</p>
<p>If <c><anno>SupName</anno>={local,Name}</c> the supervisor is registered
locally as <c>Name</c> using <c>register/2</c>. If
<c><anno>SupName</anno>={global,Name}</c> the supervisor is registered
globally as <c>Name</c> using <c>global:register_name/2</c>.
If no name is provided, the supervisor is not registered.</p>
<p><c><anno>Module</anno></c> is the name of the callback module.</p>
<p><c><anno>Args</anno></c> is an arbitrary term which is passed as
the argument to <c><anno>Module</anno>:init/1</c>.</p>
<p>If the supervisor and its child processes are successfully
created (i.e. if all child process start functions return
<c>{ok,Child}</c>, <c>{ok,Child,Info}</c>, or <c>ignore</c>)
the function returns <c>{ok,Pid}</c>, where <c>Pid</c> is
the pid of the supervisor. If there already exists a process
with the specified <c><anno>SupName</anno></c> the function returns
<c>{error,{already_started,Pid}}</c>, where <c>Pid</c> is
the pid of that process.</p>
<p>If <c><anno>Module</anno>:init/1</c> returns <c>ignore</c>, this function
returns <c>ignore</c> as well and the supervisor terminates
with reason <c>normal</c>.
If <c><anno>Module</anno>:init/1</c> fails or returns an incorrect value,
this function returns <c>{error,Term}</c> where <c>Term</c>
is a term with information about the error, and the supervisor
terminates with reason <c>Term</c>.</p>
<p>If any child process start function fails or returns an error
tuple or an erroneous value, the function returns
<c>{error,shutdown}</c> and the supervisor terminates all
started child processes and then itself with reason
<c>shutdown</c>.</p>
</desc>
</func>
<func>
<name name="start_child" arity="2"/>
<fsummary>Dynamically add a child process to a supervisor.</fsummary>
<type name="child_spec"/>
<type name="startchild_ret"/>
<type name="startchild_err"/>
<desc>
<p>Dynamically adds a child specification to the supervisor
<c><anno>SupRef</anno></c> which starts the corresponding child process.</p>
<p><marker id="SupRef"><c><anno>SupRef</anno></c></marker> can be:</p>
<list type="bulleted">
<item>the pid,</item>
<item><c>Name</c>, if the supervisor is locally registered,</item>
<item><c>{Name,Node}</c>, if the supervisor is locally
registered at another node, or</item>
<item><c>{global,Name}</c>, if the supervisor is globally
registered.</item>
</list>
<p><c><anno>ChildSpec</anno></c> should be a valid child specification
(unless the supervisor is a <c>simple_one_for_one</c>
supervisor, see below). The child process will be started by
using the start function as defined in the child
specification.</p>
<p>If the case of a <c>simple_one_for_one</c> supervisor,
the child specification defined in <c>Module:init/1</c> will
be used and <c><anno>ChildSpec</anno></c> should instead be an arbitrary
list of terms <c><anno>List</anno></c>. The child process will then be
started by appending <c><anno>List</anno></c> to the existing start
function arguments, i.e. by calling
<c>apply(M, F, A++<anno>List</anno>)</c> where <c>{M,F,A}</c> is the start
function defined in the child specification.</p>
<p>If there already exists a child specification with
the specified <c><anno>Id</anno></c>, <c><anno>ChildSpec</anno></c> is discarded and
the function returns <c>{error,already_present}</c> or
<c>{error,{already_started,<anno>Child</anno>}}</c>, depending on if
the corresponding child process is running or not.</p>
<p>If the child process start function returns <c>{ok,<anno>Child</anno>}</c>
or <c>{ok,<anno>Child</anno>,<anno>Info</anno>}</c>, the child specification and pid is
added to the supervisor and the function returns the same
value.</p>
<p>If the child process start function returns <c>ignore</c>,
the child specification is added to the supervisor, the pid
is set to <c>undefined</c> and the function returns
<c>{ok,undefined}</c>.</p>
<p>If the child process start function returns an error tuple or
an erroneous value, or if it fails, the child specification is
discarded and the function returns <c>{error,Error}</c> where
<c>Error</c> is a term containing information about the error
and child specification.</p>
</desc>
</func>
<func>
<name name="terminate_child" arity="2"/>
<fsummary>Terminate a child process belonging to a supervisor.</fsummary>
<desc>
<p>Tells the supervisor <c><anno>SupRef</anno></c> to terminate the given
child.</p>
<p>If the supervisor is not <c>simple_one_for_one</c>,
<c><anno>Id</anno></c> must be the child specification identifier. The
process, if there is one, is terminated but the child
specification is kept by the supervisor. The child process
may later be restarted by the supervisor. The child process
can also be restarted explicitly by calling
<c>restart_child/2</c>. Use <c>delete_child/2</c> to remove
the child specification.</p>
<p>If the supervisor is <c>simple_one_for_one</c>, <c><anno>Id</anno></c>
must be the child process' <c>pid()</c>. I the specified
process is alive, but is not a child of the given
supervisor, the function will return
<c>{error,not_found}</c>. If the child specification
identifier is given instead instead of a <c>pid()</c>, the
function will return <c>{error,simple_one_for_one}</c>.</p>
<p>If successful, the function returns <c>ok</c>. If there is
no child specification with the specified <c><anno>Id</anno></c>, the
function returns <c>{error,not_found}</c>.</p>
<p>See <c>start_child/2</c> for a description of
<c><anno>SupRef</anno></c>.</p>
</desc>
</func>
<func>
<name name="delete_child" arity="2"/>
<fsummary>Delete a child specification from a supervisor.</fsummary>
<desc>
<p>Tells the supervisor <c><anno>SupRef</anno></c> to delete the child
specification identified by <c><anno>Id</anno></c>. The corresponding child
process must not be running, use <c>terminate_child/2</c> to
terminate it.</p>
<p>See <seealso marker="#SupRef"><c>start_child/2</c></seealso> for a description of
<c>SupRef</c>.</p>
<p>If successful, the function returns <c>ok</c>. If the child
specification identified by <c><anno>Id</anno></c> exists but
the corresponding child process is running, the function
returns <c>{error,running}</c>. If the child specification
identified by <c><anno>Id</anno></c> does not exist, the function returns
<c>{error,not_found}</c>.</p>
</desc>
</func>
<func>
<name name="restart_child" arity="2"/>
<fsummary>Restart a terminated child process belonging to a supervisor.</fsummary>
<desc>
<p>Tells the supervisor <c><anno>SupRef</anno></c> to restart a child process
corresponding to the child specification identified by
<c><anno>Id</anno></c>. The child specification must exist and
the corresponding child process must not be running.</p>
<p>See <seealso marker="#SupRef"><c>start_child/2</c></seealso> for a description of
<c>SupRef</c>.</p>
<p>If the child specification identified by <c><anno>Id</anno></c> does not
exist, the function returns <c>{error,not_found}</c>. If
the child specification exists but the corresponding process
is already running, the function returns
<c>{error,running}</c>.</p>
<p>If the child process start function returns <c>{ok,<anno>Child</anno>}</c>
or <c>{ok,<anno>Child</anno>,<anno>Info</anno>}</c>, the pid is added to the supervisor
and the function returns the same value.</p>
<p>If the child process start function returns <c>ignore</c>,
the pid remains set to <c>undefined</c> and the function
returns <c>{ok,undefined}</c>.</p>
<p>If the child process start function returns an error tuple or
an erroneous value, or if it fails, the function returns
<c>{error,<anno>Error</anno>}</c> where <c><anno>Error</anno></c> is a term containing
information about the error.</p>
</desc>
</func>
<func>
<name name="which_children" arity="1"/>
<fsummary>Return information about all children specifications and child processes belonging to a supervisor.</fsummary>
<desc>
<p>Returns a newly created list with information about all child
specifications and child processes belonging to
the supervisor <c><anno>SupRef</anno></c>.</p>
<p>Note that calling this function when supervising a large
number of children under low memory conditions can cause an
out of memory exception.</p>
<p>See <seealso marker="#SupRef"><c>start_child/2</c></seealso> for a description of
<c>SupRef</c>.</p>
<p>The information given for each child specification/process
is:</p>
<list type="bulleted">
<item>
<p><c><anno>Id</anno></c> - as defined in the child specification or
<c>undefined</c> in the case of a
<c>simple_one_for_one</c> supervisor.</p>
</item>
<item>
<p><c><anno>Child</anno></c> - the pid of the corresponding child
process, or <c>undefined</c> if there is no such process.</p>
</item>
<item>
<p><c><anno>Type</anno></c> - as defined in the child specification.</p>
</item>
<item>
<p><c><anno>Modules</anno></c> - as defined in the child specification.</p>
</item>
</list>
</desc>
</func>
<func>
<name name="count_children" arity="1"/>
<fsummary>Return counts for the number of childspecs, active children, supervisors and workers.</fsummary>
<desc>
<p>Returns a property list (see <c>proplists</c>) containing the
counts for each of the following elements of the supervisor's
child specifications and managed processes:</p>
<list type="bulleted">
<item>
<p><c>specs</c> - the total count of children, dead or alive.</p>
</item>
<item>
<p><c>active</c> - the count of all actively running child processes
managed by this supervisor.</p>
</item>
<item>
<p><c>supervisors</c> - the count of all children marked as
child_type = supervisor in the spec list, whether or not the
child process is still alive.</p>
</item>
<item>
<p><c>workers</c> - the count of all children marked as
child_type = worker in the spec list, whether or not the child
process is still alive.</p>
</item>
</list>
</desc>
</func>
<func>
<name name="check_childspecs" arity="1"/>
<fsummary>Check if children specifications are syntactically correct.</fsummary>
<desc>
<p>This function takes a list of child specification as argument
and returns <c>ok</c> if all of them are syntactically
correct, or <c>{error,<anno>Error</anno>}</c> otherwise.</p>
</desc>
</func>
</funcs>
<section>
<title>CALLBACK FUNCTIONS</title>
<p>The following functions should be exported from a
<c>supervisor</c> callback module.</p>
</section>
<funcs>
<func>
<name>Module:init(Args) -> Result</name>
<fsummary>Return a supervisor specification.</fsummary>
<type>
<v>Args = term()</v>
<v>Result = {ok,{{RestartStrategy,MaxR,MaxT},[ChildSpec]}} | ignore</v>
<v> RestartStrategy = <seealso marker="#type-strategy">strategy()</seealso></v>
<v> MaxR = MaxT = integer()>=0</v>
<v> ChildSpec = <seealso marker="#type-child_spec">child_spec()</seealso></v>
</type>
<desc>
<p>Whenever a supervisor is started using
<c>supervisor:start_link/2,3</c>, this function is called by
the new process to find out about restart strategy, maximum
restart frequency and child specifications.</p>
<p><c>Args</c> is the <c>Args</c> argument provided to the start
function.</p>
<p><c>RestartStrategy</c> is the restart strategy and
<c>MaxR</c> and <c>MaxT</c> defines the maximum restart
frequency of the supervisor. <c>[ChildSpec]</c> is a list of
valid child specifications defining which child processes
the supervisor should start and monitor. See the discussion
about Supervision Principles above.</p>
<p>Note that when the restart strategy is
<c>simple_one_for_one</c>, the list of child specifications
must be a list with one child specification only.
(The <c>Id</c> is ignored). No child process is then started
during the initialization phase, but all children are assumed
to be started dynamically using
<c>supervisor:start_child/2</c>.</p>
<p>The function may also return <c>ignore</c>.</p>
</desc>
</func>
</funcs>
<section>
<title>SEE ALSO</title>
<p><seealso marker="gen_event">gen_event(3)</seealso>,
<seealso marker="gen_fsm">gen_fsm(3)</seealso>,
<seealso marker="gen_server">gen_server(3)</seealso>,
<seealso marker="sys">sys(3)</seealso></p>
</section>
</erlref>