The Process Component
The Process component executes commands in sub-processes.
Installation
1
$ composer require symfony/process
Note
If you install this component outside of a Symfony application, you must
require the vendor/autoload.php
file in your code to enable the class
autoloading mechanism provided by Composer. Read
this article for more details.
Usage
The Process class executes a command in a sub-process, taking care of the differences between operating system and escaping arguments to prevent security issues. It replaces PHP functions like exec, passthru, shell_exec and system:
1 2 3 4 5 6 7 8 9 10 11 12
use Symfony\Component\Process\Exception\ProcessFailedException;
use Symfony\Component\Process\Process;
$process = new Process(['ls', '-lsa']);
$process->run();
// executes after the command finishes
if (!$process->isSuccessful()) {
throw new ProcessFailedException($process);
}
echo $process->getOutput();
The getOutput()
method always returns the whole content of the standard
output of the command and getErrorOutput()
the content of the error
output. Alternatively, the getIncrementalOutput()
and getIncrementalErrorOutput()
methods return the new output since the last call.
The clearOutput() method clears the contents of the output and clearErrorOutput() clears the contents of the error output.
You can also use the Process class with the for each construct to get the output while it is generated. By default, the loop waits for new output before going to the next iteration:
1 2 3 4 5 6 7 8 9 10
$process = new Process(['ls', '-lsa']);
$process->start();
foreach ($process as $type => $data) {
if ($process::OUT === $type) {
echo "\nRead from stdout: ".$data;
} else { // $process::ERR === $type
echo "\nRead from stderr: ".$data;
}
}
Tip
The Process component internally uses a PHP iterator to get the output while
it is generated. That iterator is exposed via the getIterator()
method
to allow customizing its behavior:
1 2 3 4 5 6
$process = new Process(['ls', '-lsa']);
$process->start();
$iterator = $process->getIterator($process::ITER_SKIP_ERR | $process::ITER_KEEP_OUTPUT);
foreach ($iterator as $data) {
echo $data."\n";
}
The mustRun()
method is identical to run()
, except that it will throw
a ProcessFailedException
if the process couldn't be executed successfully (i.e. the process exited
with a non-zero code):
1 2 3 4 5 6 7 8 9 10 11 12
use Symfony\Component\Process\Exception\ProcessFailedException;
use Symfony\Component\Process\Process;
$process = new Process(['ls', '-lsa']);
try {
$process->mustRun();
echo $process->getOutput();
} catch (ProcessFailedException $exception) {
echo $exception->getMessage();
}
Tip
You can get the last output time in seconds by using the
getLastOutputTime() method.
This method returns null
if the process wasn't started!
Configuring Process Options
Symfony uses the PHP proc_open function to run the processes.
You can configure the options passed to the other_options
argument of
proc_open()
using the setOptions()
method:
1 2 3
$process = new Process(['...', '...', '...']);
// this option allows a subprocess to continue running after the main script exited
$process->setOptions(['create_new_console' => true]);
Warning
Most of the options defined by proc_open()
(such as create_new_console
and suppress_errors
) are only supported on Windows operating systems.
Check out the PHP documentation for proc_open() before using them.
Using Features From the OS Shell
Using an array of arguments is the recommended way to define commands. This saves you from any escaping and allows sending signals seamlessly (e.g. to stop processes while they run):
1 2
$process = new Process(['/path/command', '--option', 'argument', 'etc.']);
$process = new Process(['/path/to/php', '--define', 'memory_limit=1024M', '/path/to/script.php']);
If you need to use stream redirections, conditional execution, or any other feature provided by the shell of your operating system, you can also define commands as strings using the fromShellCommandline() static factory.
Each operating system provides a different syntax for their command-lines, so it becomes your responsibility to deal with escaping and portability.
When using strings to define commands, variable arguments are passed as
environment variables using the second argument of the run()
,
mustRun()
or start()
methods. Referencing them is also OS-dependent:
1 2 3 4 5 6 7 8
// On Unix-like OSes (Linux, macOS)
$process = Process::fromShellCommandline('echo "$MESSAGE"');
// On Windows
$process = Process::fromShellCommandline('echo "!MESSAGE!"');
// On both Unix-like and Windows
$process->run(null, ['MESSAGE' => 'Something to output']);
If you prefer to create portable commands that are independent from the operating system, you can write the above command as follows:
1 2
// works the same on Windows , Linux and macOS
$process = Process::fromShellCommandline('echo "${:MESSAGE}"');
Portable commands require using a syntax that is specific to the component: when
enclosing a variable name into "${:
and }"
exactly, the process object
will replace it with its escaped value, or will fail if the variable is not
found in the list of environment variables attached to the command.
Setting Environment Variables for Processes
The constructor of the Process class and
all of its methods related to executing processes (run()
, mustRun()
,
start()
, etc.) allow passing an array of environment variables to set while
running the process:
1 2 3
$process = new Process(['...'], null, ['ENV_VAR_NAME' => 'value']);
$process = Process::fromShellCommandline('...', null, ['ENV_VAR_NAME' => 'value']);
$process->run(null, ['ENV_VAR_NAME' => 'value']);
In addition to the env vars passed explicitly, processes inherit all the env
vars defined in your system. You can prevent this by setting to false
the
env vars you want to remove:
1 2 3 4
$process = new Process(['...'], null, [
'APP_ENV' => false,
'SYMFONY_DOTENV_VARS' => false,
]);
Getting real-time Process Output
When executing a long running command (like rsync
to a remote
server), you can give feedback to the end user in real-time by passing an
anonymous function to the
run() method:
1 2 3 4 5 6 7 8 9 10
use Symfony\Component\Process\Process;
$process = new Process(['ls', '-lsa']);
$process->run(function ($type, $buffer): void {
if (Process::ERR === $type) {
echo 'ERR > '.$buffer;
} else {
echo 'OUT > '.$buffer;
}
});
Note
This feature won't work as expected in servers using PHP output buffering. In those cases, either disable the output_buffering PHP option or use the ob_flush PHP function to force sending the output buffer.
Running Processes Asynchronously
You can also start the subprocess and then let it run asynchronously, retrieving output and the status in your main process whenever you need it. Use the start() method to start an asynchronous process, the isRunning() method to check if the process is done and the getOutput() method to get the output:
1 2 3 4 5 6 7 8
$process = new Process(['ls', '-lsa']);
$process->start();
while ($process->isRunning()) {
// waiting for process to finish
}
echo $process->getOutput();
You can also wait for a process to end if you started it asynchronously and are done doing other stuff:
1 2 3 4 5 6 7 8
$process = new Process(['ls', '-lsa']);
$process->start();
// ... do other things
$process->wait();
// ... do things after the process has finished
Note
The wait() method is blocking, which means that your code will halt at this line until the external process is completed.
Note
If a Response
is sent before a child process had a chance to complete,
the server process will be killed (depending on your OS). It means that
your task will be stopped right away. Running an asynchronous process
is not the same as running a process that survives its parent process.
If you want your process to survive the request/response cycle, you can
take advantage of the kernel.terminate
event, and run your command
synchronously inside this event. Be aware that kernel.terminate
is called only if you use PHP-FPM.
Danger
Beware also that if you do that, the said PHP-FPM process will not be available to serve any new request until the subprocess is finished. This means you can quickly block your FPM pool if you're not careful enough. That is why it's generally way better not to do any fancy things even after the request is sent, but to use a job queue instead.
wait() takes one optional argument: a callback that is called repeatedly whilst the process is still running, passing in the output and its type:
1 2 3 4 5 6 7 8 9 10
$process = new Process(['ls', '-lsa']);
$process->start();
$process->wait(function ($type, $buffer): void {
if (Process::ERR === $type) {
echo 'ERR > '.$buffer;
} else {
echo 'OUT > '.$buffer;
}
});
Instead of waiting until the process has finished, you can use the waitUntil() method to keep or stop waiting based on some PHP logic. The following example starts a long running process and checks its output to wait until its fully initialized:
1 2 3 4 5 6 7 8 9 10 11
$process = new Process(['/usr/bin/php', 'slow-starting-server.php']);
$process->start();
// ... do other things
// waits until the given anonymous function returns true
$process->waitUntil(function ($type, $output): bool {
return $output === 'Ready. Waiting for commands...';
});
// ... do things after the process is ready
Streaming to the Standard Input of a Process
Before a process is started, you can specify its standard input using either the
setInput() method or the 4th argument
of the constructor. The provided input can be a string, a stream resource or a
Traversable
object:
1 2 3
$process = new Process(['cat']);
$process->setInput('foobar');
$process->run();
When this input is fully written to the subprocess standard input, the corresponding pipe is closed.
In order to write to a subprocess standard input while it is running, the component provides the InputStream class:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
$input = new InputStream();
$input->write('foo');
$process = new Process(['cat']);
$process->setInput($input);
$process->start();
// ... read process output or do other things
$input->write('bar');
$input->close();
$process->wait();
// will echo: foobar
echo $process->getOutput();
The write() method accepts scalars,
stream resources or Traversable
objects as arguments. As shown in the above example,
you need to explicitly call the close()
method when you are done writing to the standard input of the subprocess.
Using PHP Streams as the Standard Input of a Process
The input of a process can also be defined using PHP streams:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
$stream = fopen('php://temporary', 'w+');
$process = new Process(['cat']);
$process->setInput($stream);
$process->start();
fwrite($stream, 'foo');
// ... read process output or do other things
fwrite($stream, 'bar');
fclose($stream);
$process->wait();
// will echo: 'foobar'
echo $process->getOutput();
Using TTY and PTY Modes
All examples above show that your program has control over the input of a
process (using setInput()
) and the output from that process (using
getOutput()
). The Process component has two special modes that tweak
the relationship between your program and the process: teletype (tty) and
pseudo-teletype (pty).
In TTY mode, you connect the input and output of the process to the input and output of your program. This allows for instance to open an editor like Vim or Nano as a process. You enable TTY mode by calling setTty():
1 2 3 4 5 6 7
$process = new Process(['vim']);
$process->setTty(true);
$process->run();
// As the output is connected to the terminal, it is no longer possible
// to read or modify the output from the process!
dump($process->getOutput()); // null
In PTY mode, your program behaves as a terminal for the process instead of a plain input and output. Some programs behave differently when interacting with a real terminal instead of another program. For instance, some programs prompt for a password when talking with a terminal. Use setPty() to enable this mode.
Stopping a Process
Any asynchronous process can be stopped at any time with the
stop() method. This method takes
two arguments: a timeout and a signal. Once the timeout is reached, the signal
is sent to the running process. The default signal sent to a process is SIGKILL
.
Please read the signal documentation below
to find out more about signal handling in the Process component:
1 2 3 4 5 6
$process = new Process(['ls', '-lsa']);
$process->start();
// ... do other things
$process->stop(3, SIGINT);
Executing PHP Code in Isolation
If you want to execute some PHP code in isolation, use the PhpProcess
instead:
1 2 3 4 5 6 7
use Symfony\Component\Process\PhpProcess;
$process = new PhpProcess(<<<EOF
<?= 'Hello World' ?>
EOF
);
$process->run();
Executing a PHP Child Process with the Same Configuration
6.4
The PhpSubprocess
helper was introduced in Symfony 6.4.
When you start a PHP process, it uses the default configuration defined in
your php.ini
file. You can bypass these options with the -d
command line
option. For example, if memory_limit
is set to 256M
, you can disable this
memory limit when running some command like this:
php -d memory_limit=-1 bin/console app:my-command
.
However, if you run the command via the Symfony Process
class, PHP will use
the settings defined in the php.ini
file. You can solve this issue by using
the PhpSubprocess class to run the command:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
use Symfony\Component\Process\Process;
class MyCommand extends Command
{
protected function execute(InputInterface $input, OutputInterface $output): int
{
// the memory_limit (and any other config option) of this command is
// the one defined in php.ini instead of the new values (optionally)
// passed via the '-d' command option
$childProcess = new Process(['bin/console', 'cache:pool:prune']);
// the memory_limit (and any other config option) of this command takes
// into account the values (optionally) passed via the '-d' command option
$childProcess = new PhpSubprocess(['bin/console', 'cache:pool:prune']);
}
}
Process Timeout
By default processes have a timeout of 60 seconds, but you can change it passing
a different timeout (in seconds) to the setTimeout()
method:
1 2 3 4 5
use Symfony\Component\Process\Process;
$process = new Process(['ls', '-lsa']);
$process->setTimeout(3600);
$process->run();
If the timeout is reached, a ProcessTimedOutException is thrown.
For long running commands, it is your responsibility to perform the timeout check regularly:
1 2 3 4 5 6 7 8 9 10 11
$process->setTimeout(3600);
$process->start();
while ($condition) {
// ...
// check if the timeout is reached
$process->checkTimeout();
usleep(200000);
}
Tip
You can get the process start time using the getStartTime()
method.
Process Idle Timeout
In contrast to the timeout of the previous paragraph, the idle timeout only considers the time since the last output was produced by the process:
1 2 3 4 5 6
use Symfony\Component\Process\Process;
$process = new Process(['something-with-variable-runtime']);
$process->setTimeout(3600);
$process->setIdleTimeout(60);
$process->run();
In the case above, a process is considered timed out, when either the total runtime exceeds 3600 seconds, or the process does not produce any output for 60 seconds.
Process Signals
When running a program asynchronously, you can send it POSIX signals with the signal() method:
1 2 3 4 5 6 7
use Symfony\Component\Process\Process;
$process = new Process(['find', '/', '-name', 'rabbit']);
$process->start();
// will send a SIGKILL to the process
$process->signal(SIGKILL);
Process Pid
You can access the pid of a running process with the getPid() method:
1 2 3 4 5 6
use Symfony\Component\Process\Process;
$process = new Process(['/usr/bin/php', 'worker.php']);
$process->start();
$pid = $process->getPid();
Disabling Output
As standard output and error output are always fetched from the underlying process, it might be convenient to disable output in some cases to save memory. Use disableOutput() and enableOutput() to toggle this feature:
1 2 3 4 5
use Symfony\Component\Process\Process;
$process = new Process(['/usr/bin/php', 'worker.php']);
$process->disableOutput();
$process->run();
Warning
You cannot enable or disable the output while the process is running.
If you disable the output, you cannot access getOutput()
,
getIncrementalOutput()
, getErrorOutput()
, getIncrementalErrorOutput()
or
setIdleTimeout()
.
However, it is possible to pass a callback to the start
, run
or mustRun
methods to handle process output in a streaming fashion.
Finding an Executable
The Process component provides a utility class called ExecutableFinder which finds and returns the absolute path of an executable:
1 2 3 4 5
use Symfony\Component\Process\ExecutableFinder;
$executableFinder = new ExecutableFinder();
$chromedriverPath = $executableFinder->find('chromedriver');
// $chromedriverPath = '/usr/local/bin/chromedriver' (the result will be different on your computer)
The find() method also takes extra parameters to specify a default value to return and extra directories where to look for the executable:
1 2 3 4
use Symfony\Component\Process\ExecutableFinder;
$executableFinder = new ExecutableFinder();
$chromedriverPath = $executableFinder->find('chromedriver', '/path/to/chromedriver', ['local-bin/']);
Finding the Executable PHP Binary
This component also provides a special utility class called PhpExecutableFinder which returns the absolute path of the executable PHP binary available on your server:
1 2 3 4 5
use Symfony\Component\Process\PhpExecutableFinder;
$phpBinaryFinder = new PhpExecutableFinder();
$phpBinaryPath = $phpBinaryFinder->find();
// $phpBinaryPath = '/usr/local/bin/php' (the result will be different on your computer)
Checking for TTY Support
Another utility provided by this component is a method called isTtySupported() which returns whether TTY is supported on the current operating system:
1 2 3
use Symfony\Component\Process\Process;
$process = (new Process())->setTty(Process::isTtySupported());