SSH failes to restart by monit

[Alexander Litvak]

I am seeing that when sshd needs to be restarted by monit. It is really successful in starting the process but for whatever reason, start command fails to generate a pid file in /var/run/. As the result monit finds process non existent and attempts so to start it again and again.

Here is my config. I attempted to output all information into a file, however the only information I see is sshd started OK.

check process sshd with pidfile "/var/run/sshd.pid"
   #start program = "/etc/init.d/sshd start"
   #stop  program = "/etc/init.d/sshd stop"
   start program = "/bin/bash -c '/etc/init.d/sshd start >>/tmp/sshd.output 2>&1'"
   stop program = "/bin/bash -c '/etc/init.d/sshd stop >>/tmp/sshd.output 2>&1'"
   if failed port 22 protocol ssh then restart
   if 5 restarts within 5 cycles then timeout

I am not sure what causes ssh script to skip writing pid file.

[Tildeslash]

Is this really a Monit problem? Maybe a permission problem or that /etc/init.d/sshd write the pid file to another location than /var/run/sshd.pid. You should also check that you can run '/etc/init.d/sshd start ' from the console as the same user you can run Monit with.

[Alexander Litvak]

Running monit in debug mode. Initiating sshd restart from monit web interface.

'sshd' restart on user request
monit daemon with PID 30584 awakened
Awakened by User defined signal 1
'sshd' trying to restart
'sshd' stop: /etc/init.d/sshd
Stopping sshd:                                             [  OK  ]
pidfile '/var/run/sshd.pid' does not exist
pidfile '/var/run/sshd.pid' does not exist
'sshd' start: /etc/init.d/sshd
pidfile '/var/run/sshd.pid' does not exist
Starting sshd:                                             [  OK  ]
'sshd' restart action done

I am confused why pid file is not written. I read that some one blamed it on sshd starting before sshd stopped but this doesn't look like the case here.

[Alexander Litvak]

Restart from console works great. When I restart file is updated. It definitely writing the file to the right place and all operated as root.. Monit was running as root in debug mode and in the init mode. SSHD starts and runs as root.

I am on Scientific Linux 6.5 which a redhat clone.

[Alexander Litvak]

Sorry for adding more confusion. It looks like using sync patch (synchronous stop , and I know it is not going to be implemented in the main code) works. I.e. I can restart the process from web and it produces sshd.pid in the right place. I am going to run more testing tonight just to make sure I am not loosing my mind.

[Tildeslash]

You can also try the new restart option which let the script control how to do restart:

check process sshd with pidfile "/var/run/sshd.pid" 
       start program = "/etc/init.d/sshd start"
       stop program = "/etc/init.d/sshd stop" 
       restart program = "/etc/init.d/sshd restart"
       if failed port 22 protocol ssh then restart 
         ...

[Alexander Litvak]

I tried restart feature and it works, at least when I issue restart from the monit webserver I have process back up with pid file updated. There are couple of minor issues

When I originally ran restart with restart command then I used to get this e-mail

PID changed Service sshd 

    Date:        Tue, 10 Jun 2014 16:04:44
    Action:      alert
    Host:       xxxxxxxxx
    Description: process PID changed from 5734 to 11238

Your faithful employee,
Monit

Now I get

PPID changed Service sshd 

    Date:        Tue, 10 Jun 2014 16:04:44
    Action:      alert
    Host:       xxxxxxx
    Description: process PPID changed from 0 to 1

Your faithful employee,
Monit

When I comment out restart command and rely on stop / start then I get no alert e-mail at all.

[Tildeslash]

Hello Alex,

the problem is fixed in the development version, you can get snapshot here: https://bitbucket.org/tildeslash/monit/get/master.tar.gz

To compile:

tar -xzf master.tar.gz
cd tildeslash*
./bootstrap
./configure
make

[Tildeslash]

fixed

[Tildeslash]

Removing version: 5.8.1 (automated comment)