Supervisord spawns only 1 instance

Question

I have the following problem:

I want to run a php script as a gearman worker, or as 5 workers to be precise. This is my supervisord.conf-addition:

[program:gearman-test-worker]
process_name=%(program_name)s_%(process_num)02d
command=/usr/bin/php /path/to/gearman-jobs/worker.php
numprocs=5
directory=/path/to/gearman-jobs/
stdout_logfile=/var/log/gearman-job-server/supervisord.log
environment=GEARMAN_USER=gearman
autostart=true
autorestart=true
user=gearman
stopsignal=KILL

When I start it supervisord -n (or in daemon mode) it will only spawn 1 instance instead of 5. The messages will look fine

[root@pc113 gearman-jobs]# supervisord -n
2013-09-03 14:24:58,775 CRIT Supervisor running as root (no user in config file)
2013-09-03 14:24:58,789 INFO /var/tmp/supervisor.sock:Medusa (V1.1.1.1) started at Tue Sep  3 14:24:58 2013
        Hostname: <unix domain socket>
        Port:/var/tmp/supervisor.sock
2013-09-03 14:24:58,850 CRIT Running without any HTTP authentication checking
2013-09-03 14:24:58,850 INFO supervisord started with pid 8722
2013-09-03 14:24:58,853 INFO spawned: 'gearman-test-worker' with pid 8723
2013-09-03 14:24:59,858 INFO success: gearman-test-worker entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

and no hint why it is only spawning a single instance. Can you help?

Answer 1

I think the problem is quite simple, the "numprocs" configuration item is only introduced in v3.0, see http://supervisord.org/configuration.html.

I was able to duplicate your problem on a stock CentOS 6.x system, which lags behind significantly in terms of releases.

# supervisord -n
2013-09-14 17:19:52,708 CRIT Supervisor running as root (no user in config file)
2013-09-14 17:19:52,724 INFO /var/tmp/supervisor.sock:Medusa (V1.1.1.1) started at Sat Sep 14 17:19:52 2013
    Hostname: <unix domain socket>
    Port:/var/tmp/supervisor.sock
2013-09-14 17:19:52,809 CRIT Running without any HTTP authentication checking
2013-09-14 17:19:52,809 INFO supervisord started with pid 21521
2013-09-14 17:19:52,817 INFO spawned: 'gearman-test-worker' with pid 21522
2013-09-14 17:19:53,820 INFO success: gearman-test-worker entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

# rpm -qa | grep supervisor
supervisor-2.1-8.el6.noarch

# uname -a
Linux dev26.sumilux.com 2.6.32-279.5.1.el6.i686 #1 SMP Tue Aug 14 21:14:52 UTC 2012 i686 i686 i386 GNU/Linux

If I upgrade the supervisor package to 3.x (actually by switching over to an ancient Fedora system), everything works as expected (I set 3 workers instead of 5).

# supervisord -n
2013-09-14 17:24:56,205 CRIT Supervisor running as root (no user in config file)
2013-09-14 17:24:56,251 INFO RPC interface 'supervisor' initialized
2013-09-14 17:24:56,252 CRIT Server 'unix_http_server' running without any HTTP authentication checking
2013-09-14 17:24:56,252 INFO supervisord started with pid 20889
2013-09-14 17:24:57,258 INFO spawned: 'gearman-test-worker_00' with pid 20890
2013-09-14 17:24:57,263 INFO spawned: 'gearman-test-worker_01' with pid 20891
2013-09-14 17:24:57,269 INFO spawned: 'gearman-test-worker_02' with pid 20892
2013-09-14 17:24:58,271 INFO success: gearman-test-worker_00 entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2013-09-14 17:24:58,271 INFO success: gearman-test-worker_01 entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2013-09-14 17:24:58,272 INFO success: gearman-test-worker_02 entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

# rpm -qa | grep supervisor
supervisor-3.0-0.5.a10.fc14.noarch