path: root/lib/observer/doc/src/ttb.xml

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

<erlref>
  <header>
    <copyright>
      <year>2002</year>
      <year>2007</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.

  The Initial Developer of the Original Code is Ericsson AB.
    </legalnotice>

    <title>ttb</title>
    <prepared>Siri hansen</prepared>
    <prepared>Bartłomiej Puzoń</prepared>
    <responsible></responsible>
    <docno>1</docno>
    <approved></approved>
    <checked></checked>
    <date>2010-08-13</date>
    <rev>PA1</rev>
    <file>ttb.sgml</file>
  </header>
  <module>ttb</module>
  <modulesummary>A base for building trace tools for distributed systems.</modulesummary>
  <description>
    <p>The Trace Tool Builder <c>ttb</c> is a base for building trace
      tools for distributed systems.
      </p>
    <p>When using <c>ttb</c>, <c>dbg</c> shall not be used in parallel.</p>
  </description>
  <funcs>
    <func>
      <name>start_trace(Nodes, Patterns, FlagSpec, Opts) -> Result</name>
      <fsummary>Start a trace port on each given node.</fsummary>
      <type>
        <v>Result = see p/2</v>
        <v>Nodes = see tracer/2</v>
        <v>Patterns = [tuple()]</v>
        <v>FlagSpec = {Procs, Flags}</v>
        <v>Proc = see p/2</v>
        <v>Flags = see p/2</v>
        <v>Opts = see tracer/2</v>
      </type>
      <desc>
        <p>This function is a shortcut allowing to start a trace with one command. Each
          tuple in <c>Patterns</c> is converted to list which is in turn passed to
          <c>ttb:tpl</c>.
          The call:<code type="none">
ttb:start_trace([Node, OtherNode],
[{mod, foo, []}, {mod, bar, 2}],
{all, call},
[{file, File}, {handler,{fun myhandler/4, S}}])</code>
          is equivalent to <code type="none">
ttb:start_trace([Node, OtherNode], [{file, File}, {handler,{fun myhandler/4, S}}]),
ttb:tpl(mod, foo, []),
ttb:tpl(mod, bar, 2, []),
ttb:p(all, call)</code>
        </p>
      </desc>
    </func>
    <func>
      <name>tracer() -> Result</name>
      <fsummary>This is equivalent to tracer(node()).</fsummary>
      <desc>
        <p>This is equivalent to <c>tracer(node())</c>.</p>
      </desc>
    </func>
    <func>
      <name>tracer(Shortcut) -> Result</name>
      <fsummary>Handy shortcuts for common tracing settings</fsummary>
      <type>
        <v>Shortcut = shell | dbg</v>
      </type>
      <desc>
        <p><c>shell</c> is equivalent to <c>tracer(node(),[{file, {local, "ttb"}}, shell])</c>.</p>
        <p><c>dbg</c> is equivalent to <c>tracer(node(),[{shell, only}])</c>.</p>
      </desc>
    </func>
    <func>
      <name>tracer(Nodes) -> Result</name>
      <fsummary>This is equivalent to tracer(Nodes,[]).</fsummary>
      <desc>
        <p>This is equivalent to <c>tracer(Nodes,[])</c>.</p>
      </desc>
    </func>
    <func>
      <name>tracer(Nodes,Opts) -> Result</name>
      <fsummary>Start a trace port on each given node.</fsummary>
      <type>
        <v>Result = {ok, ActivatedNodes} | {error,Reason}</v>
        <v>Nodes   = atom() | [atom()] | all | existing | new</v>
        <v>Opts = Opt | [Opt]</v>
        <v>Opt = {file,Client} | {handler, FormatHandler} | {process_info,PI} |
          shell | {shell, ShellSpec} | {timer, TimerSpec} | {overload, {MSec, Module, Function}}
          | {flush, MSec} | resume | {resume, FetchTimeout}</v>
        <v>TimerSpec = MSec | {MSec, StopOpts}</v>
        <v>MSec = FetchTimeout = integer()</v>
        <v>Module = Function = atom() </v>
        <v>StopOpts = see stop/2</v>
        <v>Client = File | {local, File}</v>
        <v>File = Filename | Wrap</v>
        <v>Filename = string()</v>
        <v>Wrap = {wrap,Filename} | {wrap,Filename,Size,Count}</v>
        <v>FormatHandler = See format/2</v>
        <v>PI = true | false </v>
        <v>ShellSpec = true | false | only</v>
      </type>
      <desc>
        <p>This function starts a file trace port on all given nodes
          and also points the system tracer for sequential tracing to
          the same port.
          </p>
        <p>The given <c>Filename</c> will be prefixed with the node
          name. Default <c>Filename</c> is "ttb".
          </p>
        <p><c>File={wrap,Filename,Size,Count}</c> can be used if
          the size of the trace logs must be limited. Default values are
          <c>Size=128*1024</c> and <c>Count=8</c>.
          </p>
        <p>When tracing diskless nodes, <c>ttb</c> must be started
          from an external "trace control node" with disk access, and
          <c>Client</c> must be <c>{local, File}</c>. All
          trace information is then sent to the trace control node where
          it is written to file.
          </p>
        <p>The <c>process_info</c> option indicates if process
          information should be collected. If <c>PI = true</c> (which is
          default), each process identifier <c>Pid</c> is replaced by a
          tuple <c>{Pid,ProcessInfo,Node}</c>, where <c>ProcessInfo</c>
          is the process' registered name its globally registered name,
          or its initial function. It is possible to turn off this
          functionality by setting <c>PI = false</c>.
        </p>
        <p>The <c>{shell, ShellSpec}</c> option indicates that the trace messages should
          be printed on the console as they are received by the tracing
          process. This implies <c>{local, File}</c> trace client. If the ShellSpec
          is <c>only</c> (instead of <c>true</c>), no trace logs are stored.
        </p>
        <p>The <c>shell</c> option is a shortcut for <c>{shell, true}</c>.</p>
        <p>The <c>timer</c> option indicates that the trace should be
          automatically stopped after <c>MSec</c> milliseconds. <c>StopOpts</c>
          are passed to <c>ttb:stop/2</c> command if specified (default is <c>[]</c>).
          Note that the timing is approximate, as delays related to
          network communication are always present. The timer starts after
          <c>ttb:p/2</c> is issued, so you can set up your trace patterns before.
        </p>
        <p>The <c>overload</c> option allows to enable overload
          checking on the nodes under trace. <c>Module:Function(check)</c>
          is performed each <c>MSec</c> milliseconds. If the check returns
          <c>true</c>, the tracing is disabled on a given node.<br/>
          <c>Module:Function</c> should be able to handle at least three
          atoms: <c>init</c>, <c>check</c> and <c>stop</c>. <c>init</c> and
          <c>stop</c> give the user a possibility to initialize and clean
          up the check environment.<br/>
          When a node gets overloaded, it is not possible to issue <c>ttb:p</c>
          nor any command from the <c>ttb:tp</c> family, as it would lead to
          inconsistent tracing state (different trace specifications on
          different node).
        </p>
        <p>The <c>flush</c> option periodically flushes all file trace
          port clients (see <c>dbg:flush_trace_port/1</c>). When enabled,
          the buffers are freed each <c>MSec</c> milliseconds. This option is
          not allowed with <c>{file, {local, File}}</c> tracing.
        </p>
        <p><c>{resume, FetchTimeout}</c> enables the autoresume feature.
          Whenever enabled, remote nodes try to reconnect to the controlling node
          in case they were restarted. The feature requires <c>runtime_tools</c>
          application to be started (so it has to be present in the <c>.boot</c>
          scripts if the traced nodes run with embedded erlang). If this is
          not possible, resume may be performed manually by starting
          <c>runtime_tools</c> remotely using <c>rpc:call/4</c>.<br/>
          <c>ttb</c> tries to fetch all logs from a reconnecting node before
          reinitializing the trace. This has to finish within FetchTimeout milliseconds
          or is aborted<br/>
          By default, autostart information is stored in a file called
          <c>ttb_autostart.bin</c> on each node. If this is not desired
          (i.e. on diskless nodes), a custom module to handle autostart
          information storage and retrieval can be provided by specifying
          <c>ttb_autostart_module</c> environment variable for the <c>runtime_tools</c>
          application. The module has to respond to the following API:
          <taglist>
            <tag><c>write_config(Data) -> ok</c></tag>
            <item>Store the provided data for further retrieval. It is
              important to realize that the data storage used must not
              be affected by the node crash.</item>
            <tag><c>read_config() -> {ok, Data} | {error, Error}</c></tag>
            <item>Retrieve configuration stored with <c>write_config(Data)</c>.</item>
            <tag><c>delete_config() -> ok</c></tag>
            <item>Delete configuration stored with <c>write_config(Data)</c>.
              Note that after this call any subsequent calls to <c>read_config</c>
              must return <c>{error, Error}</c>.
            </item>
          </taglist>
        </p>
        <p>The <c>resume</c> option implies the default <c>FetchTimeout</c>, which is
          10 seconds</p>
      </desc>
    </func>
    <func>
      <name>p(Procs,Flags) -> Return</name>
      <fsummary>Sets the given trace flags on the given processes.</fsummary>
      <type>
        <v>Return = {ok,[{Procs,MatchDesc}]}</v>
        <v>Procs   = Process | [Process] | all | new | existing</v>
        <v>Process = pid() | atom() | {global,atom()}</v>
        <v>Flags   = Flag | [Flag]</v>
      </type>
      <desc>
        <p>This function sets the given trace flags on the given
          processes. The <c>timestamp</c> flag is always turned on.
          </p>
        <p>Please turn to the Reference manual for module <c>dbg</c>
          for details about the possible trace flags. The parameter
          <c>MatchDesc</c> is the same as returned from <c>dbg:p/2</c></p>
        <p>Processes can be given as registered names, globally
          registered names or process identifiers. If a registered name
          is given, the flags are set on processes with this name on all
          active nodes.</p>
        <p>Issuing this command starts the timer for this trace if
          <c>timer</c> option was specified with <c>tracer/2</c>.
        </p>
      </desc>
    </func>
    <func>
      <name>tp, tpl, ctp, ctpl, ctpg</name>
      <fsummary>Set and clear trace patterns.</fsummary>
      <desc>
        <p>These functions should be used in combination with the
          <c>call</c> trace flag for setting and clearing trace
          patterns. When the <c>call</c> trace flag is set on a process,
          function calls will be traced on that process if a trace
          pattern has been set for the called function. Trace patterns
          specifies how to trace a function by using match
          specifications. Match specifications are described in the
          User's Guide for the erlang runtime system <c>erts</c>.
          </p>
        <p>These functions are equivalent to the corresponding
          functions in <c>dbg</c>, but all calls are stored in the
          history. The history buffer makes it easy to create config
          files so that the same trace environment can be setup several
          times, e.g. if you want to compare two test runs. It also
          reduces the amount of typing when using <c>ttb</c> from the
          erlang shell.
          </p>
        <taglist>
          <tag><c>tp</c></tag>
          <item>Set trace pattern on global function calls</item>
          <tag><c>tpl</c></tag>
          <item>Set trace pattern on local and global function calls</item>
          <tag><c>ctp</c></tag>
          <item>Clear trace pattern on local and global function
           calls</item>
          <tag><c>ctpl</c></tag>
          <item>Clear trace pattern on local function calls</item>
          <tag><c>ctpg</c></tag>
          <item>Clear trace pattern on global function calls</item>
        </taglist>
        <p>With <c>tp</c> and <c>tpl</c> one of match specification shortcuts
          may be used (example: <c>ttb:tp(foo_module, caller)</c>). The shortcuts are:
          <taglist>
            <item><c>return</c> - for <c>[{'_',[],[{return_trace}]}]</c>
              (report the return value)</item>
            <item><c>caller</c> - for <c>[{'_',[],[{message,{caller}}]}]</c>
              (report the calling function)</item>
            <item><c>{codestr, Str}</c> - for <c>dbg:fun2ms/1</c> arguments
              passed as strings (example: <c>"fun(_) -> return_trace() end"</c>)
            </item>
          </taglist>
        </p>
      </desc>
    </func>
    <func>
      <name>list_history() -> History</name>
      <fsummary>Returns all calls stored in history</fsummary>
      <type>
        <v>History = [{N,Func,Args}]</v>
      </type>
      <desc>
        <p>All calls to <c>ttb</c> is stored in the history. This
          function returns the current content of the history. Any entry
          can be re-executed with <c>run_history/1</c> or stored in a
          config file with <c>write_config/2/3</c>.</p>
      </desc>
    </func>
    <func>
      <name>run_history(N) -> ok | {error, Reason}</name>
      <fsummary>Executes one entry of the history</fsummary>
      <type>
        <v>N = integer() | [integer()]</v>
      </type>
      <desc>
        <p>Executes the given entry or entries from the history
          list. History can be listed with <c>list_history/0</c>.</p>
      </desc>
    </func>
    <func>
      <name>write_config(ConfigFile,Config)</name>
      <fsummary>Equivalent to write_config(ConfigFile,Config,[]).</fsummary>
      <desc>
        <p>Equivalent to <c>write_config(ConfigFile,Config,[])</c>.</p>
      </desc>
    </func>
    <func>
      <name>write_config(ConfigFile,Config,Opts) -> ok | {error,Reason}</name>
      <fsummary>Creates a config file.</fsummary>
      <type>
        <v>ConfigFile = string()</v>
        <v>Config = all | [integer()] | [{Mod,Func,Args}]</v>
        <v>Mod = atom()</v>
        <v>Func = atom()</v>
        <v>Args = [term()]</v>
        <v>Opts = Opt | [Opt]</v>
        <v>Opt = append</v>
      </type>
      <desc>
        <p>This function creates or extends a config file which can be
          used for restoring a specific configuration later.
          </p>
        <p>The content of the config file can either be fetched from
          the history or given directly as a list of
          <c>{Mod,Func,Args}</c>.
          </p>
        <p>If the complete history is to be stored in the config file
          <c>Config</c> should be <c>all</c>. If only a selected number
          of entries from the history should be stored, <c>Config</c>
          should be a list of integers pointing out the entries to be
          stored.
          </p>
        <p>If <c>Opts</c> is not given or if it is <c>[]</c>,
          <c>ConfigFile</c> is deleted and a new file is created. If
          <c>Opts = [append]</c>, <c>ConfigFile</c> will not be deleted.
          The new information will be appended at the end of the file.</p>
      </desc>
    </func>
    <func>
      <name>run_config(ConfigFile) -> ok | {error,Reason}</name>
      <fsummary>Executes all entries in a config file.</fsummary>
      <type>
        <v>ConfigFile = string()</v>
      </type>
      <desc>
        <p>Executes all entries in the given config file. Note that the history
          of the last trace is always available in the file named
          <c>ttb_last_config</c>.</p>
      </desc>
    </func>
    <func>
      <name>run_config(ConfigFile,NumList) -> ok | {error,Reason}</name>
      <fsummary>Executes selected entries from a config file.</fsummary>
      <type>
        <v>ConfigFile = string()</v>
        <v>NumList = [integer()]</v>
      </type>
      <desc>
        <p>Executes selected entries from the given config
          file. <c>NumList</c> is a list of integers pointing out the
          entries to be executed.
          </p>
        <p>The content of a config file can be listed with
          <c>list_config/1</c>.</p>
        <p> Note that the history
          of the last trace is always available in the file named
          <c>ttb_last_config</c>.</p>
      </desc>
    </func>
    <func>
      <name>list_config(ConfigFile) -> Config | {error,Reason}</name>
      <fsummary>Lists all entries in a config file.</fsummary>
      <type>
        <v>ConfigFile = string()</v>
        <v>Config = [{N,Func,Args}]</v>
      </type>
      <desc>
        <p>Lists all entries in the given config file.</p>
      </desc>
    </func>
    <func>
      <name>write_trace_info(Key,Info) -> ok</name>
      <fsummary>Writes any information to the <c>.ti</c>file.</fsummary>
      <type>
        <v>Key = term()</v>
        <v>Info = Data | fun() -> Data</v>
        <v>Data = term()</v>
      </type>
      <desc>
        <p>The <c>.ti</c> file contains <c>{Key,ValueList}</c>
          tuples. This function adds <c>Data</c> to the ValueList
          associated with <c>Key</c>. All information written with this
          function will be included in the call to the format handler.</p>
      </desc>
    </func>
    <func>
      <name>seq_trigger_ms() -> MatchSpec</name>
      <fsummary>Equivalent to seq_trigger_ms(all)</fsummary>
      <desc>
        <p>Equivalent to <c>seq_trigger_ms(all)</c></p>
      </desc>
    </func>
    <func>
      <name>seq_trigger_ms(Flags) -> MatchSpec</name>
      <fsummary>Returns a match_spec() which starts sequential tracing</fsummary>
      <type>
        <v>MatchSpec = match_spec()</v>
        <v>Flags = all | SeqTraceFlag | [SeqTraceFlag]</v>
        <v>SeqTraceFlag = atom()</v>
      </type>
      <desc>
        <p>A match specification can turn on or off sequential
          tracing. This function returns a match specification which
          turns on sequential tracing with the given <c>Flags</c>.
          </p>
        <p>This match specification can be given as the last argument
          to <c>tp</c> or <c>tpl</c>. The activated <c>Item</c> will
          then become a <em>trigger</em> for sequential tracing. This
          means that if the item is called on a process with the
          <c>call</c> trace flag set, the process will be "contaminated"
          with the seq_trace token.
          </p>
        <p>If <c>Flags = all</c>, all possible flags are set.
          </p>
        <p>Please turn to the reference manual for the
          <em><c>seq_trace</c></em> module in the <em><c>kernel</c></em>
          application to see the possible values for
          <c>SeqTraceFlag</c>.  For a description of the match_spec()
          syntax, please turn to the <em>User's guide</em> for the
          runtime system (<em>erts</em>). The chapter <em>Match Specification in Erlang</em> explains the general match
          specification "language".
          </p>
        <note>
          <p>The <em>system tracer</em> for sequential tracing is
            automatically initiated by <c>ttb</c> when a trace port is
            started with <c>ttb:tracer/0/1/2</c>.</p>
        </note>
        <p>Example of how to use the <c>seq_trigger_ms/0/1</c> function:</p>
        <code type="none">
(tiger@durin)5> ttb:tracer().
{ok,[tiger@durin]}
(tiger@durin)6> ttb:p(all,call).
{ok,{[all],[call]}}
(tiger@durin)7> ttb:tp(mod,func,ttb:seq_trigger_ms()).
{ok,[{matched,1},{saved,1}]}
(tiger@durin)8>         </code>
        <p>Whenever <c>mod:func(...)</c> is called after this, the
          seq_trace token will be set on the executing process.</p>
      </desc>
    </func>
    <func>
      <name>stop()</name>
      <fsummary>Equivalent to stop([])</fsummary>
      <desc>
        <p>Equivalent to <c>stop([])</c>.</p>
      </desc>
    </func>
    <func>
      <name>stop(Opts) -> stopped | {stopped, Dir}</name>
      <fsummary>Stop tracing and fetch/format logs from all nodes</fsummary>
      <type>
        <v>Opts = Opt | [Opt]</v>
        <v>Opt = nofetch | {fetch_dir, Dir} | format | {format, FormatOpts} | return</v>
        <v>Dir = string()</v>
        <v>FormatOpts = see format/2</v>
      </type>
      <desc>
        <p>Stops tracing on all nodes. Logs and
          trace information files are sent to the trace control
          node and stored in a directory named
          <c>ttb_upload_FileName-Timestamp</c>, where <c>Filename</c> is
          the one provided with <c>{file, File}</c> during trace setup
          and <c>Timestamp</c> is of the
          form <c>yyyymmdd-hhmmss</c>. Even logs from nodes on the same
          machine as the trace control node are moved to this directory.
          The history list is saved to a file named <c>ttb_last_config</c>
          for further reference (as it will be not longer accessible
          through history and configuration management functions (like
          <c>ttb:list_history/0</c>).
        </p>
        <p>The <c>nofetch</c> option indicates that trace logs shall not be
          collected after tracing is stopped.
        </p>
        <p>The <c>{fetch, Dir}</c> option allows to specify the directory
          to fetch the data to. If the directory already exists, an
          error is thrown.
        </p>
        <p>The <c>format</c> option indicates that the trace logs
          shall be formatted after tracing is stopped. Note that this
          fetching the logs. All logs in the fetch directory will be merged.
          You may use <c>{format, FormatOpts}</c> to pass additional
          arguments to <c>format/2</c>.</p>
        <p>The <c>return</c> option indicates that the return value
          should be <c>{stopped, Dir}</c> and not just <c>stopped</c>.
          This implies <c>fetch</c>.
        </p>
      </desc>
    </func>
    <func>
      <name>get_et_handler()</name>
      <fsummary>Returns <c>et</c> handler.</fsummary>
      <desc>
        <p>The <c>et</c> handler returned by the function may be used with <c>format/2</c>
          or <c>tracer/2</c>. Example: <c>ttb:format(Dir, [{handler, ttb:get_et_handler()}])</c>.</p>
      </desc>
    </func>
    <func>
      <name>format(File)</name>
      <fsummary>Same as <c>format(File,[])</c>.</fsummary>
      <desc>
        <p>Same as <c>format(File,[])</c>.</p>
      </desc>
    </func>
    <func>
      <name>format(File,Options) -> ok | {error, Reason}</name>
      <fsummary>Format a binary trace log</fsummary>
      <type>
        <v>File = string() | [string()]</v>
        <d>This can be the name of a binary log, a list of such logs or the name of a directory containing one or more binary logs.</d>
        <v>Options = Opt | [Opt]</v>
        <v>Opt = {out,Out} | {handler,FormatHandler} | disable_sort</v>
        <v>Out = standard_io | string()</v>
        <v>FormatHandler = {Function, InitialState}</v>
        <v>Function = fun(Fd,Trace,TraceInfo,State) -> State</v>
        <v>Fd = standard_io | FileDescriptor</v>
        <d>This is the file descriptor of the destination file <c>Out</c></d>
        <v>Trace = tuple()</v>
        <d>This is the trace message. Please turn to the Reference manual for the <c>erlang</c>module for details.</d>
        <v>TraceInfo = [{Key,ValueList}]</v>
        <d>This includes the keys <c>flags</c>, <c>client</c> and <c>node</c>,  and if <c>handler</c> is given as option to the tracer function,  this is also included. In addition all information written with the <c>write_trace_info/2</c>function is included. </d>
      </type>
      <desc>
        <p>Reads the given binary trace log(s). The logs are processed
          in the order of their timestamp as long as <c>disable_sort</c>
          option is not given.
        </p>
        <p>If <c>FormatHandler = {Function,InitialState}</c>,
          <c>Function</c> will be called for each trace message. If
          <c>FormatHandler = get_et_handler()</c>, <c>et_viewer</c> in
          the <em>Event Tracer</em> application (<c>et</c>) is used for presenting
          the trace log graphically. <c>ttb</c> provides a few different
          filters which can be selected from the Filter menu in the
          <c>et_viewer</c>. If <c>FormatHandler</c> is not given, a
          default handler is used which presents each trace message as a
          line of text.
          </p>
        <p>If <c>Out</c> is given, <c>FormatHandler</c> gets the
          file descriptor to <c>Out</c> as the first parameter.
          </p>
        <p><c>Out</c> is ignored if <c>et</c> format handler is used.
          </p>
        <p>Wrap logs can be formatted one by one or all in one go. To
          format one of the wrap logs in a set, give the exact name of
          the file. To format the whole set of wrap logs, give the name
          with '*' instead of the wrap count. See examples in the
          <c>ttb</c> User's Guide.</p>
      </desc>
    </func>
  </funcs>
</erlref>