Re: Process being restarted by monit validate even if it is running

[Martin Pala]

You can use pattern based check for processes which don't have pidfile (no need 
to create wrapper which will write the pidfile):

        check process v2api matching "/usr/local/bin/node 
/usr/local/share/nodeapis/server.js"
                ...

The "as uid root" should be removed from start+stop script (as you described, 
Monit is running as root already).

For debugging, stop Monit and and start it again in foreground ("-I" = capital 
"i") with -v option:

        monit -vI

Monit will log details about each test to console.

Regards,
Martin



On 07 Mar 2014, at 12:21, Mehul Ved <address@hidden> wrote:

> # ls -ld /var/run
> lrwxrwxrwx 1 root root 4 Jul  5  2013 /var/run -> /run
> 
> # ls -ld /var/run/node/
> drwxr-xr-x 2 root root 120 Mar  7 08:52 /var/run/node/
> 
> # ls -l /var/run/node/v2.pid 
> -rw-r--r-- 1 root root 4 Mar  7 08:52 /var/run/node/v2.pid
> 
> # lsof -i :2812
> COMMAND   PID USER   FD   TYPE   DEVICE SIZE/OFF NODE NAME
> monit   14382 root    6u  IPv4 70768563      0t0  TCP localhost:2812 (LISTEN)
> 
> I am running monit as root and also monit validate is run from root login. 
> 
> I had added "as uid root" specifically because I was having problem reading 
> pidfile without that. Probably redundant now.
> Another thing I should have made clear. This is only happening only for 
> processes that I have written a start/stop script. My script gets the pid of 
> the process and echo's it into the given file. To the best of my knowledge 
> that's the correct thing to do with a pidfile. And considering that monit is 
> getting the pid correctly, I believe that part is fine. It even manages to 
> stop the program correctly, which would have failed if the program wasn't 
> running with correct pid. 
> 
> I am failing to figure out as to why monit doesn't see that process running. 
> Any suggested debugging steps I can follow?
> 
> ________________________________________
> From: address@hidden <address@hidden> on behalf of Martin Pala 
> <address@hidden>
> Sent: Friday, March 07, 2014 4:26 PM
> To: This is the general mailing list for monit
> Subject: Re: Process being restarted by monit validate even if it is running
> 
> Hi,
> 
> is Monit running as root or as different user?
> 
> If it is running as root, then the "as uid root" in stop/start programs is 
> not necessary:
> 
>        start program = "/usr/local/bin/nodeinit v2 start" as uid root
> 
> If it is running as different user (which may be the reason for adding "as 
> uid root", but that most probably won't work, as the user won't have 
> permission to switch to root unless the binary is setuid or sudo is used), 
> then it is possible that Monit cannot read the pidfile, please check the 
> permissions of the whole path to the pidfile and the pidfile itself:
> 
>        ls -ld /var/run
>        ls -ld /var/run/node
>        ls -l /var/run/node/v2.pid
> 
> You can run Monit in debug mode to get more details about the test progress:
> 
>        monit -vI
> 
> 
> Regards,
> Martin
> 
> 
> On 07 Mar 2014, at 10:03, Mehul Ved <address@hidden> wrote:
> 
>> Hi,
>>  I have a process which I am monitoring with following rules
>> 
>> check process v2api with pidfile /var/run/node/v2.pid
>>  start program = "/usr/local/bin/nodeinit v2 start"
>>    as uid root
>>  stop program = "/usr/local/bin/nodeinit v2 stop"
>>    as uid root
>>  if failed host 127.0.0.1 port 10400 protocol http
>>    request /api/v2/ping
>>    with timeout 10 seconds
>>    then restart
>>  if 5 restarts within 10 cycles then alert
>> 
>> Before running monit validate, I checked the pidfile of the process
>> 
>> # grep [0-9]* /var/run/node/
>> /var/run/node/v2.pid:31566
>> 
>> # cat /var/run/node/v2.pid
>> 31566
>> 
>> I also verified with ps on the process id
>> 
>> # ps aux | grep 31566
>> root     31566  0.3  1.7 605684 29820 ?        Sl   05:24   0:01 
>> /usr/local/bin/node /usr/local/share/nodeapis/server.js
>> 
>> 
>> Now when I run
>> 
>> # monit validate --verbose
>> 'v2api' Error testing process id [31566] -- No such process
>> 'v2api' process is not running
>> Does not exist notification is sent to address@hidden
>> 'v2api' trying to restart
>> 'v2api' stop: /usr/local/bin/nodeinit
>> 'v2api' Error testing process id [31566] -- No such process
>> /usr/local/bin/nodeinit: line 80: kill: (31566) - No such process
>> Killed v2 process with pid 31566
>> monit: pidfile '/var/run/node/v2.pid' does not exist
>> monit: pidfile '/var/run/node/v2.pid' does not exist
>> 'v2api' start: /usr/local/bin/nodeinit
>> monit: pidfile '/var/run/node/v2.pid' does not exist
>> v2 has started with PID: 1150
>> 
>> It complains that the process is not running.
>> 
>> I am using the development version of monit that Martin linked to a couple 
>> of days back, with websocket support.
>> 
>> # monit --version
>> This is Monit version 5.8
>> Copyright (C) 2001-2014 Tildeslash Ltd. All Rights Reserved.
>> 
>> ​
>> 
>> 
>> 
>> 
>> 
>> 
>> --
>> To unsubscribe:
>> https://lists.nongnu.org/mailman/listinfo/monit-general
> 
> 
> --
> To unsubscribe:
> https://lists.nongnu.org/mailman/listinfo/monit-general
> --
> To unsubscribe:
> https://lists.nongnu.org/mailman/listinfo/monit-general