Supervisor

Supervisor is a client/server system that allows its users to control a number of processes on UNIX-like operating systems. It was inspired by the following:

- It is often inconvenient to need to write "rc.d" scripts for every single process instance. rc.d scripts are a great lowest-common-denominator form of process initialization/autostart/management, but they can be painful to write and maintain. Additionally, rc.d scripts cannot automatically restart a crashed process and many programs do not restart themselves properly on a crash. Supervisord starts processes as its subprocesses, and can be configured to automatically restart them on a crash. It can also automatically be configured to start processes on its own invocation.
- It's often difficult to get accurate up/down status on processes on UNIX. Pidfiles often lie. Supervisord starts processes as subprocesses, so it always knows the true up/down status of its children and can be queried conveniently for this data.
- Users who need to control process state often need only to do that. They don't want or need full-blown shell access to the machine on which the processes are running. Supervisorctl allows a very limited form of access to the machine, essentially allowing users to see process status and control supervisord-controlled subprocesses by emitting "stop", "start", and "restart" commands from a simple shell or web UI.
- Users often need to control processes on many machines. Supervisor provides a simple, secure, and uniform mechanism for interactively and automatically controlling processes on groups of machines.
- Processes which listen on "low" TCP ports often need to be started and restarted as the root user (a UNIX misfeature). It's usually the case that it's perfectly fine to allow "normal" people to stop or restart such a process, but providing them with shell access is often impractical, and providing them with root access or sudo access is often impossible. It's also (rightly) difficult to explain to them why this problem exists. If supervisord is started as root, it is possible to allow "normal" users to control such processes without needing to explain the intricacies of the problem to them.
- Processes often need to be started and stopped in groups, sometimes even in a "priority order". It's often difficult to explain to people how to do this. Supervisor allows you to assign priorities to processes, and allows user to emit commands via the supervisorctl client like "start all", and "restart all", which starts them in the preassigned priority order. Additionally, processes can be grouped into "process groups" and a set of logically related processes can be stopped and started as a unit.otification system was added.

What is new in this release:

• The behavior of the program option user has changed. In all previous versions, if supervisord failed to switch to the user, a warning would be sent to the stderr log but the child process would still be spawned. This means that a mistake in the config file could result in a child process being unintentionally spawned as root. Now, supervisord will not spawn the child unless it was able to successfully switch to the user. Thanks to Igor Partola for reporting this issue.
• If a user specified in the config file does not exist on the system, supervisord will now print an error and refuse to start.
• Reverted a change to logging introduced in 3.0b1 that was intended to allow multiple processes to log to the same file with the rotating log handler. The implementation caused supervisord to crash during reload and to leak file handles. Also, since log rotation options are given on a per-program basis, impossible configurations could be created (conflicting rotation options for the same file). Given this and that supervisord now has syslog support, it was decided to remove this feature. A warning was added to the documentation that two processes may not log to the same file.
• Fixed a bug where parsing command= could cause supervisord to crash if shlex.split() fails, such as a bad quoting. Patch by Scott Wilson.
• It is now possible to use supervisorctl on a machine with no supervisord.conf file by supplying the connection information in command line options. Patch by Jens Rantil.
• Fixed a bug where supervisord would crash if the syslog handler was used and supervisord received SIGUSR2 (log reopen request).
• Fixed an XML-RPC bug where calling supervisor.getProcessInfo() with a bad name would cause a 500 Internal Server Error rather than the returning a BAD_NAME fault.
• Added a favicon to the web interface. Patch by Caio Ariede.
• Fixed a test failure due to incorrect handling of daylight savings time in the childutils tests. Patch by Ildar Hizbulin.
• Fixed a number of pyflakes warnings for unused variables, imports, and dead code. Patch by Philippe Ombredanne.

What is new in version 3.0 Beta 1:

• Fixed a bug where parsing environment= did not verify that key/value pairs were correctly separated. Patch by Martijn Pieters.
• Fixed a bug in the HTTP server code that could cause unnecessary delays when sending large responses. Patch by Philip Zeyliger.
• When supervisord starts up as root, if the -c flag was not provided, a warning is now emitted to the console. Rationale: supervisord looks in the current working directory for a supervisord.conf file; someone might trick the root user into starting supervisord while cd'ed into a directory that has a rogue supervisord.conf.
• A warning was added to the documentation about the security implications of starting supervisord without the -c flag.
• Add a boolean program option stopasgroup, defaulting to false. When true, the flag causes supervisor to send the stop signal to the whole process group. This is useful for programs, such as Flask in debug mode, that do not propagate stop signals to their children, leaving them orphaned.
• Python 2.3 is no longer supported. The last version that supported Python 2.3 is Supervisor 3.0a12.
• Removed the unused "supervisor_rpc" entry point from setup.py.
• Fixed a bug in the rotating log handler that would cause unexpected results when two processes were set to log to the same file. Patch by Whit Morriss.
• Fixed a bug in config file reloading where each reload could leak memory because a list of warning messages would be appended but never cleared. Patch by Philip Zeyliger.
• Added a new Syslog log handler. Thanks to Denis Bilenko, Nathan L. Smith, and Jason R. Coombs, who each contributed to the patch.
• Put all change history into a single file (CHANGES.txt).

What is new in version 3.0 Alpha 12:

• Released to replace a broken 3.0a11 package where non-Python files were not included in the package.

What is new in version 3.0 Alpha 10:

• Fixed the stylesheet of the web interface so the footer line won't overlap a long process list. Thanks to Derek DeVries for the patch.
• Allow rpc interface plugins to register new events types.
• Bug fix for FCGI sockets not getting cleaned up when the reload command is issued from supervisorctl. Also, the default behavior has changed for FCGI sockets. They are new closed whenever the number ofrunning processes in a group hits zero. Previously, the sockets were kept open unless a group-level stop command was issued.
• Better error message when HTTP server cannot reverse-resolve a hostname to an IP address. Previous behavior: show a socket error. Current behavior: spit out a suggestion to stdout.
• Environment variables set via environment= value within
• [supervisord] section had no effect. Thanks to Wyatt Baldwin for a patch.
• Fix bug where stopping process would cause process output that happened after the stop request was issued to be lost. See https://github.com/Supervisor/supervisor/issues/11.
• Moved 2.X change log entries into HISTORY.txt.
• Converted CHANGES.txt and README.txt into proper ReStructuredText and included them in the long_description in setup.py.
• Added a tox.ini to the package (run via tox in the package dir). Tests supervisor on multiple Python versions.

Requirements:

• Python