Using supervisor as CRON

Question

Is there any way to configure supervisor to run some command every X seconds(like CRON)?

I see example with eventlistener and TICK_ event

[eventlistener:memmon]
command=memmon -a 200MB -m bob@example.com
events=TICK_60

But it runs the command only once.

Answer 1

Why invent the wheel? You can use cron and supervisord together.

In supervisord, create a task with autostart=false

In cron, use * * * * * supervisorctl start <taskname> to start the task every minute

Answer 2

Problem

As you are seeing in the memmon example, supervisord is not executing memmon -a 200MB -m bob@example.com at each event. Rather, it is starting this event listener once (or potentially a few times if you configure a pool) and then sends each new event over standard input of an existing process.

Solution

Consequently, you really need to find or write a supervisor compatible event listener for each additional type of capability you want to trigger on events.

Example Method of Implementation

Setup the configuration and write a listener

write a supervisord.cfg event section

[eventlistener:passthru]
command=/tmp/simple.py /bin/date -u +"%%s %%S:%%H:%%d:%%m"
events=TICK_60

(note- the escaping of % for configParser )

write a simple.py Event Listener

Create this simple.py listener by making changes to the example listener from the docs so it executes its first argument with any remaining arguments:

#! /usr/bin/python
import sys
import subprocess

def write_stdout(s):
    sys.stdout.write(s)
    sys.stdout.flush()

def write_stderr(s):
    sys.stderr.write(s)
    sys.stderr.flush()

def main(args):
    while 1:
        write_stdout('READY\n') # transition from ACKNOWLEDGED to READY
        line = sys.stdin.readline()  # read header line from stdin
        write_stderr(line) # print it out to stderr
        headers = dict([ x.split(':') for x in line.split() ])
        data = sys.stdin.read(int(headers['len'])) # read the event payload
        res = subprocess.call(args, stdout=sys.stderr); # don't mess with real stdout
        write_stderr(data)
        write_stdout('RESULT 2\nOK') # transition from READY to ACKNOWLEDGED

if __name__ == '__main__':
    main(sys.argv[1:])
    import sys

Make sure the supervisor configuration works

$ supervisorctl [-c cfg]
supervisor> status
passthru                         RUNNING   pid 4471, uptime 0:00:32
supervisor> tail passthru
  OKREADY
  RESULT 2
  OKREADY
  ...
supervisor> tail passthru stderr
supervisor> tail passthru stderr
  ver:3.0 server:supervisor serial:0 pool:passthru poolserial:0 eventname:TICK_60 len:15
  1451411161 01:17:29:12 <--- output
  when:1451411160ver:3.0 server:supervisor serial:1 pool:passthru poolserial:1 eventname:TICK_60 len:15
  1451411220 00:17:29:12 <--- output
  when:1451411220

Now date -u +"%s %S:%H:%d:%m" is running every 60 seconds.

Swapping in the desired command

create an executable script

/tmp/hiworld.php:

#! /usr/bin/php
<?= "hiya\n"; 

(chmod +x ...)

change the listener's arguments in supervisord.cfg

[eventlistener:passthru]
command=/tmp/simple.py /tmp/hiworld.php
;stdout_logfile=/tmp/passthru 
events=TICK_60
;autorestart=true
;startsecs=0

reload supervisord and test (reread seems not to detect this change)

supervisor> reload
   Really restart the remote supervisord process y/N? y
   Restarted supervisord
supervisor> status
   passthru                         RUNNING   pid 6017, uptime 0:00:10
supervisor> tail passthru stderr
supervisor> status
   passthru                         RUNNING   pid 6017, uptime 0:00:21
supervisor> status
   passthru                         RUNNING   pid 6017, uptime 0:01:01
supervisor> tail passthru stderr
   ver:3.0 server:supervisor serial:316 pool:passthru poolserial:0 eventname:TICK_60 len:15
    hiya
   when:1418926740
supervisor> 

End

Now the desired command is running every 60 seconds. You are now read to adjust particulars of permissions, locations, logs, etc.

Answer 3

Supervisor does not support this easily.

But to achieve your goal you can just use supervisor to start cron (for a docker container for instance):

https://gist.github.com/martinrusev/7015e393d46647dbad15

Install cron in your docker (apt-get install cron with a debian-like docker)

In supervisor config:

[program:cron]
command=cron -f -L 15
autostart=true
autorestart=true

-f is for foreground, -L 15 is to have all cron logs output.

Then use a user crontab, the global /etc/crontab or any crontab special directories (/etc/cron.hourly, /etc/cron.daily, …)

Answer 4

You can call bash sleep command:

[program:mycmd]
command=bash -c 'sleep 300 && exec <your command here>'
autorestart=true

This will run your command every 5 minutes. Do not forget exec part to replace bash process with your command so supervisor will get correct exit code.

Answer 5

You can use crobtab to manage and schedule your supervisor programs.

Use command supervisorctl start <program_name>

Note: This will start only single instance of supervisor program. If it's already running and crontab tries to trigger it again, supervisorctl start command will not start a new instance.