Re: [monit] alert restarts service

[Martin Pala]

gorlist wrote:
> Thanks for the reply.
> I understand.
>
> 1. Whenever I stop it manually though it only sends a "Does not exist
> Service sshd" message and the date. It does not follow the default
> more detailed format that ends with "Your faithful employee, monit".
> Why?

The message which ends with "Your faithful employee, monit" is the 
default mail format. It is possible to override the default template 
with "mail-format" option. See monit manual for more details.

Maybe you used this option ?

It could be good if we can see the configuration.



> 2. What would be a good way to implement a 'hacky' remote host service
> restart? Like quick and dirty fix.

It could be possible to call the actions on remote hosts using remote 
monit web interface hooked to start/stop/exec action with wget for example.


> 3. I have a problem when sending mails. I use postfix, I have it
> configured to "Internet site" and sometimes it does not let monit send
> the alert mail. The mail command will succeed with all hosts (3/3).
> When monit does it, it only succeeds with gmail.com (1/3). The other
> two is yahoo.com and my own mailserver.

Please can you provide the network trace of the communication with the 
failed mailservers?




> 4. How is monit allowed to start and stop services, since AFAIK only
> root is privileged to do so.

Monit has the same privileges as the user under which monit is running. 
When monit is running as root, it's possible to setuid/setgid when 
executing start/stop/exec actions to different user/group.



Martin



> On Fri, Apr 18, 2008 at 1:22 AM, Martin Pala <address@hidden> wrote:
> > Hi,
> >
> >  the restart action when monitored process exits is implicit.
> >
> >  We plan to add options to allow to override the default behavior:
> >  http://www.tildeslash.com/monit/doc/next.php#36
> >
> >  Since the process existence comes before socket testing, in your case monit
> > detects that ssh is not running and restarts it. The socket/port related
> > alert action was not even evaluated (since it makes no sense when process is
> > not running).
> >
> >  If you want to cleanly stop process which is monitored by monit, you should
> > do "monit stop ssh", so monit will perform the stop (otherwise when the
> > process exists out of monit control, monit tries to fix the problem).
> >
> >
> >  Remote hosts can be monitored using "check host" statement ... currently it
> > is not possible to restart remote services (just alerts). The possibility to
> > restart remote services will be part of M/Monit.
> >
> >
> >  We don't have IRC channel for support currently.
> >
> >
> >  Cheers,
> >  Martin
> >
> >
> >
> >
> >
> >  gorlist wrote:
> >
> > > Hello,
> > >
> > > firstly let me say a big thank you to the developers for such a wonderful
> > tool.
> > > So my monitrc looks like this:
> > >  check process sshd with pidfile /var/run/sshd.pid
> > >   start program  "/etc/init.d/sshd start"
> > >   stop program  "/etc/init.d/sshd stop"
> > >   if failed port 22 protocol ssh then alert
> > >
> > > The thing is that if I stop the sshd manually, the next cycle it
> > > displays "Does not exist", sends an email to the global address and
> > > after that it starts the daemon.
> > > That means 2 cycles after I stop the sshd, it starts it even if I
> > > haven't specified restart (I have only alert).
> > > Why does it do that?
> > >
> > > Also, is there a way that I can monitor daemons on remote hosts and
> > > restart them if needed? If yes, how?
> > >
> > > Finally, is there an IRC channel or something except this list to get
> > support?
> > > Thank you
> > >
> > >
> > > --
> > > To unsubscribe:
> > > http://lists.nongnu.org/mailman/listinfo/monit-general
> >
> >  --
> >  To unsubscribe:
> >  http://lists.nongnu.org/mailman/listinfo/monit-general
>
>
> --
> To unsubscribe:
> http://lists.nongnu.org/mailman/listinfo/monit-general