Multiple processes started when explicit start commands are used on dependant processes.

[P Holdaway]

Hi.
  
I am controlling monit through a PHP web interface and issue start and stop
commands to monit on an arbitrary set of processes.

  I am seeing processes which are actually running but become unmonitored.

  The scripts I use make some effort to run one and only one instance of a
process and check that the process ID actually corresponds to the monitored
process.

  In an effort to track down this problem is I think the following may be
the root cause...

#######################################################

  There appears to be a race condition when issuing monit start commands on
processes that are dependant on each other.

  To illustrate...

  Given processes b1 to b8 are dependant on process a.

  If I stop all processes and then issue a separate start command for each
process, I see in monit.log that the process b1 is being started twice. Once
as a result of the dependency on process a, and once because of an explicit
start command.

  I have attached some files to help reproduce this problem.

  As an aside note that process "a" changes its status in the web interface
from "running" to "Changed" unnecessarily? I have made the poll time 60
seconds to illustrate this.

########

[ Jan 27 10:59:55] Monitoring disabled -- service b8
[ Jan 27 10:59:55] 'b8' stop: /root/monit/run.sh
[ Jan 27 10:59:56] Monitoring disabled -- service b7
[ Jan 27 10:59:56] 'b7' stop: /root/monit/run.sh
[ Jan 27 10:59:57] Monitoring disabled -- service b6
[ Jan 27 10:59:57] 'b6' stop: /root/monit/run.sh
[ Jan 27 10:59:59] Monitoring disabled -- service b5
[ Jan 27 10:59:59] 'b5' stop: /root/monit/run.sh
[ Jan 27 11:00:00] Monitoring disabled -- service b4
[ Jan 27 11:00:00] 'b4' stop: /root/monit/run.sh
[ Jan 27 11:00:01] Monitoring disabled -- service b3
[ Jan 27 11:00:01] 'b3' stop: /root/monit/run.sh
[ Jan 27 11:00:02] Monitoring disabled -- service b2
[ Jan 27 11:00:02] 'b2' stop: /root/monit/run.sh
[ Jan 27 11:00:03] Monitoring disabled -- service b1
[ Jan 27 11:00:03] 'b1' stop: /root/monit/run.sh
[ Jan 27 11:00:04] Monitoring disabled -- service a
[ Jan 27 11:00:04] 'a' stop: /root/monit/run.sh
[ Jan 27 11:00:05] 'a' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service a
[ Jan 27 11:00:05] 'b8' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b8
[ Jan 27 11:00:05] 'b7' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b7
[ Jan 27 11:00:05] 'b6' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b6
[ Jan 27 11:00:05] 'b5' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b5
[ Jan 27 11:00:05] 'b4' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b4
[ Jan 27 11:00:05] 'b3' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b3
[ Jan 27 11:00:05] 'b2' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b2
[ Jan 27 11:00:05] 'b1' start: /root/monit/run.sh
[ Jan 27 11:00:05] Monitoring enabled -- service b1
[ Jan 27 11:00:05] 'b1' start: /root/monit/run.sh
[ Jan 27 11:00:05] monit: Process already running -- process b2
[ Jan 27 11:00:05] monit: Process already running -- process b3
[ Jan 27 11:00:05] monit: Process already running -- process b4
[ Jan 27 11:00:05] monit: Process already running -- process b5
[ Jan 27 11:00:05] monit: Process already running -- process b6
[ Jan 27 11:00:05] monit: Process already running -- process b7
[ Jan 27 11:00:05] monit: Process already running -- process b8

########

Regards,

  Peter Holdaway


--
Peter Holdaway                            TechnoCom Corporation
Phone: 760 438 5115 ext 132               2030 Corte del Nogal, Suite 200
Fax:   760 438 5815                       Carlsbad, CA 92011
Email: address@hidden   http://www.technocom-wireless.com/

monitrc
Description: Binary data

restart_all.sh
Description: Binary data

run.sh
Description: Binary data