OS Process Manager for Erlang VM

Version: 1.9-182-g237c04a

Authors: Serge Aleynikov (saleyn(at)gmail.com).

Erlexec - OS Process Manager for the Erlang VM


Execute and control OS processes from Erlang/OTP.

This project implements an Erlang application with a C++ port program that gives light-weight Erlang processes fine-grain control over execution of OS processes.

The following features are supported:

This application provides significantly better control over OS processes than built-in erlang:open_port/2 command with a {spawn, Command} option, and performs proper OS child process cleanup when the emulator exits.

The erlexec application has been in production use by Erlang and Elixir systems, and is stable.


If you find this project useful, please donate to:

Supported Platforms

Linux, Solaris, FreeBSD, OpenBSD, MacOS X


See http://saleyn.github.io/erlexec


Erlang: import as a dependency

 [% ...
  {erlexec, "~> 1.0"}
    % ...

Elixir: import as a dependency

defp deps do
    # ...
    {:erlexec, "~> 1.0"}


Make sure you have rebar or rebar3 installed locally and the rebar script is in the path.

If you are deploying the application on Linux and would like to take advantage of exec-port running tasks using effective user IDs different from the real user ID that started exec-port, then either make sure that libcap-dev[el] library is installed or make sure that the user running the port program has sudo rights.

OS-specific libcap-dev installation instructions:

$ git clone git@github.com:saleyn/erlexec.git
$ make
# NOTE: for disabling optimized build of exec-port, do the following instead:
$ OPTIMIZE=0 make

By default port program's implementation uses poll(2) call for event demultiplexing. If you prefer to use select(2), set the following environment variable:

$ USE_POLL=0 make


The program is distributed under BSD license.

Copyright @copy; 2003 Serge Aleynikov


  |   +----+ +----+ +----+    |
  |   |Pid1| |Pid2| |PidN|    |   Erlang light-weight Pids associated
  |   +----+ +----+ +----+    |   one-to-one with managed OsPids
  |         \   |   /         |
  |          \  |  /          |
  |           \ | / (links)   |
  |         +------+          |
  |         | exec |          |   Exec application running in Erlang VM
  |         +------+          |
  | Erlang VM   |             |
          | exec-port |           Port program (separate OS process)
           /    |    \
 (optional stdin/stdout/stderr pipes)
         /      |      \
    +------+ +------+ +------+
    |OsPid1| |OsPid2| |OsPidN|    Managed Child OS processes
    +------+ +------+ +------+

Configuration Options

See description of types in exec:exec_options().

The exec-port program requires the SHELL variable to be set. If you are running Erlang inside a docker container, you might need to ensure that SHELL is properly set prior to starting the emulator.


Starting/stopping an OS process

1> exec:start().                                        % Start the port program.
2> {ok, _, I} = exec:run_link("sleep 1000", []).        % Run a shell command to sleep for 1000s.
3> exec:stop(I).                                        % Kill the shell command.
ok                                                      % Note that this could also be accomplished
                                                        % by doing exec:stop(pid(0,34,0)).

In Elixir:

iex(1)> :exec.start
{:ok, #PID<0.112.0>}
iex(2)> :exec.run("echo ok", [:sync, :stdout])
{:ok, [stdout: ["ok\n"]]}

Clearing environment or unsetting an env variable of the child process

%% Clear environment with {env, [clear]} option:
10> f(Bin), {ok, [{stdout, [Bin]}]} = exec:run("env", [sync, stdout, {env, [clear]}]), p(re:split(Bin, <<"\n">>)).
[<<"PWD=/home/...">>,<<"SHLVL=0">>, <<"_=/usr/bin/env">>,<<>>]
%% Clear env and add a "TEST" env variable:
11> f(Bin), {ok, [{stdout, [Bin]}]} = exec:run("env", [sync, stdout, {env, [clear, {"TEST", "xxx"}]}]), p(re:split(Bin, <<"\n">>)).
[<<"PWD=/home/...">>,<<"SHLVL=0">>, <<"_=/usr/bin/env">>,<<"TEST=xxx">>,<<>>]
%% Unset an "EMU" env variable:
11> f(Bin), {ok, [{stdout, [Bin]}]} = exec:run("env", [sync, stdout, {env, [{"EMU", false}]}]), p(re:split(Bin, <<"\n">>)).

Running exec-port as another effective user

In order to be able to use this feature the current user must either have sudo rights or the exec-port file must be owned by root and have the SUID bit set (use: chown root:root exec-port; chmod 4555 exec-port):

$ ll priv/x86_64-unknown-linux-gnu/exec-port
-rwsr-xr-x 1 root root 777336 Dec  8 10:02 ./priv/x86_64-unknown-linux-gnu/exec-port

If the effective user doesn't have rights to access the exec-port program in the real user's directory, then the exec-port can be copied to some shared location, which will be specified at startup using {portexe, "/path/to/exec-port"}.

$ cp $(find . -name exec-port) /tmp
$ chmod 755 /tmp/exec-port
$ whoami
$ erl
1> exec:start([{user, "wheel"}, {portexe, "/tmp/exec-port"}]).  % Start the port program as effective user "wheel".
$ ps haxo user,comm | grep exec-port
wheel      exec-port

Allowing exec-port to run commands as other effective users

In order to be able to use this feature the current user must either have sudo rights or the exec-port file must have the SUID bit set, and the exec-port file must have the capabilities set as described in the "Build" section above.

The port program will initially be started as root, and then it will switch the effective user to {user, User} and set process capabilities to cap_setuid,cap_kill,cap_sys_nice. After that it'll allow to run child programs under effective users listed in the {limit_users, Users} option.

$ whoami
$ erl
1> Opts = [root, {user, "wheel"}, {limit_users, ["alex","guest"]}],
2> exec:start(Opts).                                    % Start the port program as effective user "wheel"
                                                        % and allow it to execute commands as "alex" or "guest".
3> exec:run("whoami", [sync, stdout, {user, "alex"}]).  % Command is executed under effective user "alex"
$ ps haxo user,comm | grep exec-port
wheel      exec-port

Running the port program as root

While running the port program as root is highly discouraged, since it opens a security hole that gives users an ability to damage the system, for those who might need such an option, here is how to get it done (PROCEED AT YOUR OWN RISK!!!).

Note: in this case exec would use sudo exec-port to run it as root or the exec-port must have the SUID bit set (4555) and be owned by root. The other (DANGEROUS and firmly DISCOURAGED!!!) alternative is to run erl as root:

$ whoami
# Make sure the exec-port can run as root:
$ sudo _build/default/lib/erlexec/priv/*/exec-port --whoami
$ erl
1> exec:start([root, {user, "root"}, {limit_users, ["root"]}]).
2> exec:run("whoami", [sync, stdout]).
{ok, [{stdout, [<<"root\n">>]}]}
$ ps haxo user,comm | grep exec-port
root       exec-port

Killing an OS process

Note that killing a process can be accomplished by running kill(3) command in an external shell, or by executing exec:kill/2.

1> f(I), {ok, _, I} = exec:run_link("sleep 1000", []).
2> exec:kill(I, 15).
** exception error: {exit_status,15}                    % Our shell died because we linked to the
                                                        % killed shell process via exec:run_link/2.
3> exec:status(15).                                     % Examine the exit status.
{signal,15,false}                                       % The program got SIGTERM signal and produced
                                                        % no core file.

Using a custom success return code

1> exec:start_link([]).
2> exec:run_link("sleep 1", [{success_exit_code, 0}, sync]).
3> exec:run("sleep 1", [{success_exit_code, 1}, sync]).
{error,[{exit_status,1}]}                               % Note that the command returns exit code 1

Redirecting OS process stdout to a file

7> f(I), {ok, _, I} = exec:run_link("for i in 1 2 3; do echo \"Test$i\"; done",
    [{stdout, "/tmp/output"}]).
8> io:format("~s", [binary_to_list(element(2, file:read_file("/tmp/output")))]),

Redirecting OS process stdout to screen, an Erlang process or a custom function

9> exec:run("echo Test", [{stdout, print}]).
Got stdout from 29651: <<"Test\n">>
10> exec:run("for i in 1 2 3; do sleep 1; echo \"Iter$i\"; done",
            [{stdout, fun(S,OsPid,D) -> io:format("Got ~w from ~w: ~p\n", [S,OsPid,D]) end}]).
Got stdout from 29652: <<"Iter1\n">>
Got stdout from 29652: <<"Iter2\n">>
Got stdout from 29652: <<"Iter3\n">>
% Note that stdout/stderr options are equivanet to {stdout, self()}, {stderr, self()} 
11> exec:run("echo Hello World!; echo ERR!! 1>&2", [stdout, stderr]).
12> flush().
Shell got {stdout,18382,<<"Hello World!\n">>}
Shell got {stderr,18382,<<"ERR!!\n">>}

Appending OS process stdout to a file

13> exec:run("for i in 1 2 3; do echo TEST$i; done",
        [{stdout, "/tmp/out", [append, {mode, 8#600}]}, sync]),
14> exec:run("echo Test4; done", [{stdout, "/tmp/out", [append, {mode, 8#600}]}, sync]),
15> file:delete("/tmp/out").

Setting up a monitor for the OS process

> f(I), f(P), {ok, P, I} = exec:run("echo ok", [{stdout, self()}, monitor]).
16> flush().                                                                  
Shell got {stdout,18950,<<"ok\n">>}
Shell got {'DOWN',18950,process,<0.263.0>,normal}

Managing an externally started OS process

This command allows to instruct erlexec to begin monitoring given OS process and notify Erlang when the process exits. It is also able to send signals to the process and kill it.

% Start an externally managed OS process and retrieve its OS PID:
17> spawn(fun() -> os:cmd("echo $$ > /tmp/pid; sleep 15") end).
18> f(P), P = list_to_integer(lists:reverse(tl(lists:reverse(binary_to_list(element(2,
% Manage the process and get notified by a monitor when it exits:
19> exec:manage(P, [monitor]).
% Wait for monitor notification
20> f(M), receive M -> M end.
21> file:delete("/tmp/pid").

Specifying a custom process shutdown delay in seconds

% Execute an OS process (script) that blocks SIGTERM with custom kill timeout, and monitor
22> f(I), {ok, _, I} = exec:run("trap '' SIGTERM; sleep 30", [{kill_timeout, 3}, monitor]).
% Attempt to stop the OS process
23> exec:stop(I).
% Wait for its completion
24> f(M), receive M -> M after 10000 -> timeout end.                                          

Specifying a custom kill command for a process

% Execute an OS process (script) that blocks SIGTERM, and uses a custom kill command,
% which kills it with a SIGINT. Add a monitor so that we can wait for process exit
% notification. Note the use of the special environment variable "CHILD_PID" by the
% kill command. This environment variable is set by the port program before invoking
% the kill command:
2> f(I), {ok, _, I} = exec:run("trap '' SIGTERM; sleep 30", [{kill, "kill -n 2 ${CHILD_PID}"},
                                                             {kill_timeout, 2}, monitor]).
% Try to kill by SIGTERM. This does nothing, since the process is blocking SIGTERM:
3> exec:kill(I, sigterm), f(M), receive M -> M after 0 -> timeout end.
% Attempt to stop the OS process
4> exec:stop(I).
% Wait for its completion
5> f(M), receive M -> M after 1000 -> timeout end.                                          

Communicating with an OS process via STDIN

% Execute an OS process (script) that reads STDIN and echoes it back to Erlang
25> f(I), {ok, _, I} = exec:run("read x; echo \"Got: $x\"", [stdin, stdout, monitor]).
% Send the OS process some data via its stdin
26> exec:send(I, <<"Test data\n">>).                                                  
% Get the response written to processes stdout
27> f(M), receive M -> M after 10000 -> timeout end.
{stdout,26431,<<"Got: Test data\n">>}
% Confirm that the process exited
28> f(M), receive M -> M after 10000 -> timeout end.

Communicating with an OS process via STDIN and sending end-of-file

2> Watcher = spawn(fun F() -> receive Msg -> io:format("Got: ~p\n", [Msg]), F() after 60000 -> ok end end).
3> f(Pid), f(OsPid), {ok, Pid, OsPid} = exec:run("tac", [stdin, {stdout, Watcher}, {stderr, Watcher}]).
4> exec:send(Pid, <<"foo\n">>).
5> exec:send(Pid, <<"bar\n">>).
6> exec:send(Pid, <<"baz\n">>).
7> exec:send(Pid, eof).
Got: {stdout,26143,<<"baz\nbar\nfoo\n">>}

Running OS commands synchronously

% Execute an shell script that blocks for 1 second and return its termination code
29> exec:run("sleep 1; echo Test", [sync]).
% By default all I/O is redirected to /dev/null, so no output is captured
% 'stdout' option instructs the port program to capture stdout and return it to caller
30> exec:run("sleep 1; echo Test", [stdout, sync]).
{ok,[{stdout, [<<"Test\n">>]}]}
% Execute a non-existing command
31> exec:run("echo1 Test", [sync, stdout, stderr]).   
        {stderr,[<<"/bin/bash: echo1: command not found\n">>]}]}
% Capture stdout/stderr of the executed command
32> exec:run("echo Test; echo Err 1>&2", [sync, stdout, stderr]).    
% Redirect stderr to stdout
33> exec:run("echo Test 1>&2", [{stderr, stdout}, stdout, sync]).
{ok, [{stdout, [<<"Test\n">>]}]}

Running OS commands with/without shell

% Execute a command by an OS shell interpreter
34> exec:run("/bin/echo ok", [sync, stdout]).
{ok, [{stdout, [<<"ok\n">>]}]}
% Execute an executable without a shell
35> exec:run(["/bin/echo", "ok"], [sync, stdout])).
{ok, [{stdout, [<<"ok\n">>]}]}
% Execute a shell with custom options
36> exec:run(["/bin/bash", "-c", "echo ok"], [sync, stdout])).
{ok, [{stdout, [<<"ok\n">>]}]}

Running OS commands with pseudo terminal (pty)

% Execute a command without a pty
37> exec:run("echo hello", [sync, stdout]).
{ok, [{stdout,[<<"hello\n">>]}]}
% Execute a command with a pty
38> exec:run("echo hello", [sync, stdout, pty]).

Kill a process group at process exit

% In the following scenario the process P0 will create a new process group
% equal to the OS pid of that process (value = GID). The next two commands
% are assigned to the same process group GID. As soon as the P0 process exits
% P1 and P2 will also get terminated by signal 15 (SIGTERM):
39> {ok, P0, GID} = exec:run("sleep 10",  [{group, 0},   kill_group]).
40> {ok, P1,   _} = exec:run("sleep 15",  [{group, GID}, monitor]).
41> {ok, P2,   _} = exec:run("sleep 15",  [{group, GID}, monitor]).
42> flush().
Shell got {'DOWN',25307,process,<0.39.0>,{exit_status,15}}
Shell got {'DOWN',25308,process,<0.41.0>,{exit_status,15}}