Help with action methods

[Alexander Litvak]

I am trying to use monit in our organization and I got confused by few things.  Can one of you help me?

 

1.       I wanted to use check  program  to check a status of existing  application

 

So I placed this entry

 

check program app with path "/home/alex/bin/app_ctl status"

         start program = "/home/alex/bin/app_ctl start" as uid alex and gid alex

         stop program = "/home/alex/bin/app_ctl stop" as uid alex and gid alex

        if status != 0 then restart

 

Monit detected my application being down but then said that Stop, start, or restart methods are not defined for process check app

 

2.      This app listens on sip port so I wanted to check for response and restart if it fails.

 

check host myhost with address "myhost"

        start program = "/home/alex/bin/app_ctl start" as uid alex and gid alex

         stop program = "/home/alex/bin/app_ctl stop" as uid alex and gid alex

if failed

               port 5060 type UDP protocol SIP

with target "myhost:5060" and maxforward 1

then restart

                             

Monit detected my application being unresponsive but then said that “Stop, start, or restart methods are not defined for process check myhost”

 

I see that if I use check process with pid | match   action stop / start  methods do work.   But documents are not clear on the scope of the action methods, at least it is not clear to me why alert, monitor, exec would work and stop / start / restart not in this case.  I am sure I am doing something wrong here but I am new to this project.

 

I tried using exec instead of start restart .  My app_ctl can do restart as well.  But I noticed that for some reason this is not very reliable as exec cycles through multiple restarts until the whole thing settles down.

 

I am doing something like

 

              

check program app with path "/home/alex/bin/app_ctl status”

        if status != 0 then exec  = "/home/alex/bin/app_ctl restart" as uid alex and gid alex

 

It seems to me that negative status is being retained in the next cycle.   It keeps restarting the process at least several times.  I ran exec with start instead of the restart and after it successfully started process on the first cycle it attempted to start it again which my app_ctl objected to with message

 

App is still running please stop it first...

 

Two cycles later I see the messages (running in console mode for debugging)

 

monit: Cannot open proc file /proc/15247/stat -- No such file or directory

system statistic error -- cannot read /proc/15247/stat

monit: Cannot open proc file /proc/15248/stat -- No such file or directory

system statistic error -- cannot read /proc/15248/stat

monit: Cannot open proc file /proc/15249/stat -- No such file or directory

system statistic error -- cannot read /proc/15249/stat

 

               Needless to say I have tested my app_ctl script and  running it from shell I can stop start and restart app without any issues.  It properly detects PID and status of app.

 

               Sorry for a long post, monit got me excited because it presented many interesting opportunities and has a long history.  Any help would be appreciated.

 

Alex