[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [monit] monitoring nfs mounts

From: Martin Pala
Subject: Re: [monit] monitoring nfs mounts
Date: Fri, 7 Dec 2007 23:15:33 +0100

Monit does restart action - this will call stop->start. The stop method allows to cleanup the environment after the process crashed, which may be necessary in some cases.

When monit calls start method, it checks whether the process is running. In your case the start script in "stop program" started the service already. When the "start program" was called, it checked whether the process is running and since it was running, the start method retuned as there was no action needed.

The web interface allows manual control for the services - the user is responsible for his own action (if you request monit to call just "start" method, it will do it). The restart is monit's reaction to given event (process disappeared) which consists of stop&start.


On Dec 7, 2007, at 10:49 PM, Teresa Havel wrote:

So when monit detects a process not running, it uses a "restart" and not just a "start" for its method of automatically starting the process? I see by the browser interface, a person has the option of using a "start" or "restart". So when I manually used the "start" button it worked, but monit could not automatically start is because it uses "restart". It is strange that since I am now using my start script as a dummy in the stop method, that I don't have the processes being started twice. But I will use the dummy stop method you suggested just to be safe.

Martin Pala wrote:
Monit restart action requires the stop method, since it calls the stop method first and then start. You can add dummy method like:

  stop program = "/bin/true"


Teresa Havel wrote:
I did check the logs and it did show the error when the process was not running. Then in the next line it would say:
monit: Start or stop method not defined -- process tcpaprscvt
It would not start automatically. But it would work when I used the start service button on the browser interface. Based on the log message, I began to wonder if monit needed to have a stop method present in monitrc, even though it only needed to run the start method. I added a stop method, which was just the same method that I used in start, because I figured it wasn't going to have to use it anyway, it only uses the stop method if you select stop or restart. After I did that, it began to work properly, automatically starting the process when it determined it had stopped. I belileve that solves my problem for now. Thanks for the suggestions.

Martin Pala wrote:
You can run monit in verbose mode with -v options and check the monit logs. When the file is not available, then monit should report error.

You can also check the system logs and/or mnttab using monit's content test and check remote services using tcp/udp tests. You can also use hint no. 13 from monit's FAQ: http://


Leng Siakkhasone wrote:
I have a mail server that mount nfs mounts from a file server. I would like to be notified if/when the file server is unreachable by the mail server. I tried using a token file with the following:

# /homes/home01 is the nfs mount
check file home01 with path /homes/home01/monit.token
   group hyper

But that did not work. 'monit summary' still showed it as accessible even though the file server was offline.

I would appreciate any suggestions and/or strategies for monitoring an nfs mount. Thanks.


To unsubscribe:
To unsubscribe:

To unsubscribe: general<teresa_m_havel.vcf>
To unsubscribe:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]