Hello,
I am trying to configure monit to monitor "php-resque",
https://github.com/chrisboulton/php-resque/tree/v1.1 . php-resque is a library
to create background jobs. You can run a single worker to N, it does the job
well.
The problem: I can't monitor at all when it runs with N processes. Internally,
while php-resque is executed will try to fork a process, if it's not running it
will create a new one at when atleast one worker is still alive. It auto-prune
dead workers and start new ones. This means that PIDs usually changes, now I've
this running:
PID 6028 running
PID 6029 running
PID 6030 running
The problem is that the way php-resque works, if PID 6028 is dead, it will
restart php-resque. This means killing all the PIDs and starting it. This isn't
why I expected because if there is atleast one running PID it will always start
all the missing ones.
Situation (I've 3 workers running):
PID 6028 not found running
PID 6029 running
PID 6030 running
monit see that 6028 is not running, so it restarts:
PID 6029 running
PID 6030 running
PID 13570 running
PID 13571 running
PID 13572 running
When 6029 is found not running, it will do the same and add 3 more workers
(becuase is what does /etc/init.d/php-resque start).
I tried to search an example of similar situations but I don't find anything.
What will be cool is to be able to check if all PIDs are dead and not only the
first one.
I see that you can check not only for process but for files: "check file
apache_rc with path /etc/init.d/apache" so I assume that there is a way to
using for example a bash script that will return "OK" or "ERROR" or something
similar to let know monit what to do. I can't find documentation on this so I'm
a bit lost.
Thanks in advance,
Isern
--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general
