Paste Script


Ian Bicking <>







Paste Script is being maintained on life support. That means that critical bugs will be fixed, and support for new versions of Python will be handled, but other than that new features are not being considered. Development has moved to GitHub.


If you are developer, see the Developer Documentation; this will tell you how to add commands and templates to paster. For a list of updates see the news file.

Paste Script is released under the MIT license.


Paste Script has passed version 1.0. Paste Script is in maintenance mode. Bugs will be fixed, new features are not being considered.

paster create

This creates the skeleton for new projects. Many different kinds of projects have created skeletons for their projects (Pylons, TurboGears, ZopeSkel, and others).

For a tutorial for making new skeletons, see this tutorial from Lucas Szybalski. It also discusses creating new subcommands for paster.

paster serve

The one useful command you may want to know about for paster is paster serve. This serves an application described in a Paste Deploy configuration file.


A quickstart (and example), if not complete explanation:

use = egg:PasteEnabledPackage
option1 = foo
option2 = bar

use = egg:PasteScript#wsgiutils
host =
port = 80

egg:PasteEnabledPackage refers to some package that has been prepared for use with paste.deploy, and options given to that package. If you are starting out with some framework, you’ll have to reference some documentation for that framework to paste-deploy-ify your application (or read the paste.deploy documentation).

In the same file is a server description. egg:PasteScript#wsgiutils is a server (named wsgiutils) provided by this package, based on WSGIUtils. And we pass various options particular to that server.

Other packages can provide servers, but currently Paste Script includes glue for these:


A SimpleHTTPServer based threaded HTTP server, using WSGIUtils.


This set of servers supports SCGI, FastCGI and AJP protocols, for connection an external web server (like Apache) to your application. Both threaded and forking versions are available. This is based on flup.

There is the start of support for twisted.web2 in paste.script.twisted_web2_server; patches welcome.

Running the Server

paster serve --help gives useful output:

usage: /usr/local/bin/paster serve [options] CONFIG_FILE [start|stop|restart|status]
Serve the described application

If start/stop/restart is given, then it will start (normal
operation), stop (--stop-daemon), or do both.  You probably want
``--daemon`` as well for stopping.

  -h, --help            show this help message and exit
  -v, --verbose
  -q, --quiet
  -nNAME, --app-name=NAME
                        Load the named application (default main)
                        Use the named server.
                        Use the named server as defined in the configuration
                        file (default: main)
  --daemon              Run in daemon (background) mode
  --pid-file=FILENAME   Save PID to file (default to if running in
                        daemon mode)
  --log-file=LOG_FILE   Save output to the given log file (redirects stdout)
  --reload              Use auto-restart file monitor
                        Seconds between checking files (low number can cause
                        significant CPU usage)
  --status              Show the status of the (presumably daemonized) server
  --user=USERNAME       Set the user (usually only possible when run as root)
  --group=GROUP         Set the group (usually only possible when run as root)
  --stop-daemon         Stop a daemonized server (given a PID file, or default

Basically you give it a configuration file. If you don’t do anything else, it’ll serve the [app:main] application with the [server:main] server. You can pass in --server-name=foo to serve the [server:foo] section (or even --server-name=config:foo.ini to use a separate configuration file).

Similarly you can use --app-name=foo to serve [app:foo].

--daemon will run the server in the backgroup, --user and --group will set the user, as you might want to do from a start script (run as root). If you don’t give a --pid-file it will write the pid to (in the current directory).

--stop-daemon will stop the daemon in or whatever --pid-file you give. --log-file will redirect stdout and stderr to that file.

--reload will start the reload monitor, and restart the server whenever a file is edited. This can be a little expensive, but is very useful during development.

#! Scripts

On Posix (Linux, Unix, etc) systems you can turn your configuration files into executable scripts.

First make the file executable (chmod +x config_file.ini). The you should add a line like this to the top of the file:

#!/usr/bin/env paster

You can include a command and command-line options in an [exe] section, like:

command = serve
daemon = true
user = nobody
group = nobody

(use true and false for options that don’t take an argument).

If you use daemon = true then you’ll be able to use the script as an rc script, so you can do:

$ sudo ./config_file.ini start
$ sudo ./config_file.ini restart

and so forth.

Note that this is a little wonky still on some platforms and shells (notably it doesn’t work under csh). If you get an error about “Command config_file.ini not known” then this probably won’t work for you. In the future an additional script to paster will be added just for this purpose.