pm2 is a process manager for Node apps with a builtin load-balancer.
pm2 is perfect when you need to spread your stateless NodeJS code accross all CPUs available on a server, to keep all processes alive forever and to 0s reload them.
- Builtin load-balancer (using the native cluster module)
- Script daemonization
- 0s downtime reload for Node
- Startup scripts for Ubuntu/CentOS (use updaterc.d for Ubuntu and chkconfig for others)
- Stop unstable process (avoid infinite loop)
- Monitoring in console
- HTTP API
- Remote control and real time interface API
Tested with Node v0.11, v0.10, v0.8 (https://travis-ci.org/Unitech/pm2). Compatible with CoffeeScript. Works on Linux & MacOS.
We gonna release a very nice product, a dashboard to monitor every part of your NodeJS applications. Here are some links :
- Pitch + Survey People who fill the survey will be eligible for free license
- Newsletter Subscribe to be kept informed
We are also looking for AngularJS developers and designers contact us at contact AT unitech DOT io
Thanks in advance and we hope that you like pm2 !
- Installation
- Usage/Features
- Tutorial : How To Use PM2 to Setup a Node.js Production Environment
- Pid file, error and out Log files
- Different ways to launch a process
- 0s downtime reload
- CoffeeScript
- Enabling Harmony
- Accept JSON app configuration via pipe from standard input
- Is my production server ready for PM2
- Listing processes : pm2 list
- Monitoring processes (CPU/RAM) : pm2 monit
- Startup script generation : pm2 startup
- Log aggregation : pm2 logs
- Fork mode
- Customization
- API health end point : pm2 web
- JSON processes declaration
- Contributing/Development mode
- Known bugs
- Launching the tests
- They talk about it
- License
npm install pm2@latest -g
$ npm install pm2@latest -g # Install pm2 command line globally
$ pm2 start app.js -i 4 # Daemonize pm2 and Start 4 clustered instances of app.js
# You can also pass the 'max' params to start
# the right numbers of processes depending of CPUs
$ pm2 start app.js --name my-api # Name process
$ pm2 start app.js --no-daemon # Don't daemonize pm2
$ pm2 list # Display all processes status
$ pm2 monit # Monitor all processes
$ pm2 logs # Display all processes logs in streaming
$ pm2 flush # Empty all log file
$ pm2 stop all # Stop all processes
$ pm2 restart all # Restart all processes
$ pm2 reload all # Will 0s downtime reload (for NETWORKED processes)
$ pm2 stop 0 # Stop specific process id
$ pm2 restart 0 # Restart specific process id
$ pm2 delete 0 # Will remove process from pm2 list
$ pm2 delete all # Will remove all processes from pm2 list
$ pm2 ping # Ensure pm2 dameon has been launched
$ pm2 startup ubuntu # Generate init script for ubuntu to keep processes alive on restart
# ubuntu/centos
$ pm2 web # Launch Health computer API endpoint (http://localhost:9615)
$ pm2 dump # Backup current processes managed by pm2
$ pm2 resurrect # Restore backup
$ pm2 sendSignal SIGUSR2 signal.js # Send system signal to script
For other nature scripts :
$ pm2 start echo.php
$ pm2 start echo.py
$ pm2 start echo.sh
$ pm2 start echo.rb
$ pm2 start echo.pl
$ pm2 start app.js -i max # Will start maximum processes depending on available CPUs
$ pm2 start app.js -i 3 # Will start 3 processes
$ pm2 start app.js --node-args="--debug=7001 --trace-deprecation" # --node-args command line option to pass options to node
$ pm2 start app.js -x # Start app.js in fork mode instead of cluster
$ pm2 start app.js -x -- -a 23 # Start app.js in fork mode and pass arguments (-a 23)
$ pm2 start app.js --name serverone # Start a process an name it as server one
# you can now stop the process by doing
# pm2 stop serverone
$ pm2 start app.json # Start processes with options declared in app.json
# Go to chapter Multi process JSON declaration for more
$ pm2 start app.js -i max -- -a 23 # Pass arguments after -- to app.js
$ pm2 start app.js -i max -e err.log -o out.log # Will start and generate a configuration file
You can also execute app in other languages (the fork mode):
$ pm2 start my-bash-script.sh -x --interpreter bash
$ pm2 start my-python-script.py -x --interpreter python
How To Use PM2 to Setup a Node.js Production Environment On An Ubuntu VPS
## Pid file, error and out Log filesBy default every logs (error and out), pids files, dump, pm2 logs are located in ~/.pm2/
.pm2/
├── dump.pm2
├── custom_options.sh
├── pm2.log
├── pm2.pid
├── logs
└── pids
This feature permits to reload code without losing in process connections. Works for apps in cluster_mode (the default mode) that uses sockets (express or other).
$ pm2 reload all
$ pm2 reload my-api
Thanks to TruongSinh Tran-Nguyen https://github.com/truongsinh
$ pm2 gracefulReload all
Instead of just processing remaining connections, gracefulReload
will also send a shutdown
message to your process, so you can close all database/socket.io/* connections and be sure that your process will properly exit.
process.on('message', function(msg) {
if (msg == 'shutdown') {
// Your process is going to be reloaded
// Close all database/socket.io/* connections
console.log('Closing all connections...');
setTimeout(function() {
console.log('Finished closing connections');
// You can exit to faster the process or it will be
// automatically killed after 4000ms.
// You can override the timeout by modifying PM2_GRACEFUL_TIMEOUT
process.exit(0);
}, 1500);
}
});
var http = require('http');
http.createServer(function(req, res) {
res.writeHead(200);
res.end('hey');
}).listen(8000, function() {
console.log('listening');
});
$ pm2 start my_app.coffee
That's all !
## Enabling Harmony ES6You can enable Harmony ES6 by setting PM2_NODE_OPTIONS='--harmony'
environment variable option when you start pm2 (pm2 should not be already daemonized).
To pass this option by default, you can edit ~/.pm2/custom_options.sh
and add :
export PM2_NODE_OPTIONS='--harmony'
Then :
$ pm2 dump
$ pm2 exit
$ pm2 resurrect
If ES6 has been enabled you should see this message at the beggining of each pm2 commands :
● ES6 mode
$ pm2 start my_app.js --node-args="--harmony"
The default mode of PM2 consists of wrapping the code of your node app into the Node Cluster module. It's called the cluster mode. There is also a more classical way to execute your app, like node-forever does, called the fork mode.
In fork mode almost all options are the same as the cluster mode. But no reload, gracefulReload.
By using the fork mode you will lose core features of PM2 like the automatic clusterization of your code over all CPUs available and the 0s reload.
So use it if you only need a forever-like behaviour.
Here is how to start your app in fork :
$ pm2 start app.js -x # Will start your app.js in fork mode
$ pm2 list # You will see that on the row "mode" it's written "fork"
You can also exec scripts in other languages :
$ pm2 start my-bash-script.sh -x --interpreter bash
$ pm2 start my-python-script.py -x --interpreter python
PR :
#!/bin/bash
read -d '' my_json <<_EOF_
[{
"name" : "app1",
"script" : "/home/projects/pm2_nodetest/app.js",
"instances" : "4",
"error_file" : "./logz/child-err.log",
"out_file" : "./logz/child-out.log",
"pid_file" : "./logz/child.pid",
"exec_mode" : "cluster_mode",
"port" : 4200
}]
_EOF_
echo $my_json | pm2 start -
Just try the tests before using PM2 on your production server
$ git clone https://github.com/Unitech/pm2.git
$ cd pm2
$ npm install # Or do NODE_ENV=development npm install if some packages are missing
$ npm test
If a test is broken please report us issues here Also make sure you have all dependencies needed. For Ubuntu :
$ sudo apt-get install build-essential
$ wget -qO- https://raw.github.com/creationix/nvm/master/install.sh | sh
$ nvm install v0.11.9
$ nvm use v0.11.9
$ nvm alias default v0.11.9
List infos about all processes managed by pm2. It shows also how many times a process has been restarted because of an unhandled exception.
Monitor CPU and memory usage of every node process (and also clustered processes) managed by pm2.
PM2 provides an automatic way to keep Node processes alive on server restart. On exit it will dump the process list and their environment and will resurrect them on startup.
It uses System V init script for Ubuntu/CentOS/Redhat (maybe it works on other sys but not 100% sure).
$ pm2 startup ubuntu # then follow the command instruction
$ pm2 startup centos # will use chkconfig instead of updaterc.d
$ pm2 startup redhat # not very stable for redhat
Init script generated are located in /etc/init.d/pm2-init.sh.
Use this for systemd distributions, like Fedora and Archlinux. Note that all running services must be stopped and your old dump will be overwritten.
$ pm2 kill
$ pm2 startup systemd
This will create, enable and start the service.
The -u username
option permits to specify which user has to start the process at startup.
NOTE that this user must have access to npm, apps and node ! So the best way is to log with this user su -l www
, then do pm2 startup -u www
.
Internally it uses sudo -u $USER
.
Display logs in streaming of all processes, without having to do a tail -f or something else. You can also pass [name|id] as parameter to stream only the log of a specified process.
PM2 can disserve an API endpoint to monitor processes and computer health (cpu usage, memory, network interfaces...)
pm2 web
You can edit these options by editing the file ~/.pm2/custom_options.sh
These variables can be customized :
DAEMON_BIND_HOST : process.env.PM2_BIND_ADDR || 'localhost',
DAEMON_RPC_PORT : process.env.PM2_RPC_PORT || 6666, // RPC commands
DAEMON_PUB_PORT : process.env.PM2_PUB_PORT || 6667, // Realtime events
DEBUG : process.env.PM2_DEBUG || false,
WEB_INTERFACE : process.env.PM2_API_PORT || 9615,
GRACEFUL_TIMEOUT : parseInt(process.env.PM2_GRACEFUL_TIMEOUT) || 4000,
PM2_NODE_OPTIONS : ''
processes.json :
[{
"name" : "echo",
"script" : "./examples/args.js",
"args" : "['--toto=heya coco', '-d', '1']",
"env": {
"NODE_ENV": "production",
"AWESOME_SERVICE_API_TOKEN": "xxx"
}
}
,{
"name" : "api",
"script" : "./examples/child.js",
"instances" : "4",
"error_file" : "./examples/child-err.log",
"out_file" : "./examples/child-out.log",
"pid_file" : "./examples/child.pid",
"exec_mode" : "cluster_mode",
"port" : 9005
},{
"min_uptime" : "100",
"name" : "auto-kill",
"exec_mode" : "fork_mode",
"script" : "./examples/killfast.js"
}]
Then with the cli :
$ pm2 start processes.json
$ pm2 stop processes.json
$ pm2 delete processes.json
$ pm2 restart processes.json
To hack PM2, it's pretty simple :
$ pm2 kill # kill the current pm2
$ git clone my_pm2_fork.git
$ cd pm2/
$ DEBUG=* PM2_DEBUG=true ./bin/pm2 --no-daemon
Each time you edit the code be sure to restart pm2 to make changes taking effect.
$ npm install git://github.com/Unitech/pm2#development -g
- Remove init script :
sudo update-rc.d -f pm2-init.sh remove
First, install the lastest pm2 version :
$ npm install -g pm2@latest
- Node 0.10.x doesn't free script port when stopped. It's due to the NodeJS cluster module. So if you feel that this problem is important for your use case, use the fork mode the fork mode instead. By using the fork mode you will lose core features of PM2 like the automatic clusterization of your code over all CPUs available and the 0s reload.
$ pm2 start index.js -x # start my app in fork mode
For more informations about this issue : #74
Cannot read property 'getsockname' of undefined
When using the cluster mode (by default) you can't use ports from 0 to 1024. If you really need to exec in this range use the fork mode with the -x
parameter.
By using the fork mode you will lose core features of PM2 like the automatic clusterization of your code over all CPUs available and the 0s reload.
npm test
- http://devo.ps/blog/2013/06/26/goodbye-node-forever-hello-pm2.html
- https://coderwall.com/p/igdqyw
- http://revdancatt.com/2013/09/17/node-day-1-getting-the-server-installing-node-and-pm2/
- https://medium.com/tech-talk/e7c0b0e5ce3c
- Clusterize your Node networked script without adding one line of code
- Fully tested
- Monitor process/cluster processes health (status, memory, cpu usage, restarted time) via CLI (htop like)
- Monitor server health (processes, cpu core...) via JSON api (pm2 web)
- Launch multiple applications via JSON
- Forever keep alive processes
- Log streaming in realtime (pm2 logs)
- Log uncaught exceptions in error logs
- Track restarted time
- Auto stop processes who exit too fast
- Dump current processes and resurrect (upstart)
- Remote administration/status checking
- Builtin Inter process communication channel (message bus)
- Auto start of the script at start (upstart)
- V8 GC memory leak detection
- Web interface
- Keeping monitoring data
- Integrated wrk utils endpoint benchmark
Thanks to Devo.ps and Wiredcraft for their knowledge and expertise.
# LicenseFiles in lib/ are made available under the terms of the GNU Affero General Public License (AGPL). pm2-interface is made under the terms of the Apache V2 license.