[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Program status testing - alert contains stdout and not stderr
|
From: |
Jan-Henrik Haukeland |
|
Subject: |
Re: Program status testing - alert contains stdout and not stderr |
|
Date: |
Tue, 16 Jul 2013 22:44:05 +0200 |
On 15 Jul 2013, at 14:36, Nestor Urquiza <address@hidden> wrote:
> (apparently the timeout is matching the cycle time) no matter the fact that
> I added the timeoit:
> check program myscript with path "/usr/myuser/myscript" with timeout 1800
> seconds
This is correct, timeout is not exact, but depends on your poll cycle. Monit
store the time at program start and for each poll cycle, Monit test if timeout
has been reached (start time + timeout > now ). This means that Monit can be
between cycles when the program is expected to timeout, but on the next Monit
cycle, timeout will be handled.
> It looks like monit will kill all myscript processes when the timeout occurs
> instead of just the correct pid.
Monit kills the /usr/myuser/myscript process on timeout. If this process starts
other processes they might also be killed depending on how these processes are
started. I.e. if they was forked and demonised etc.
> Is this issue related to a bug?
Not really, but the current program test has some caveats. Please read this
part of the manual to understand the “shortcomings",
http://mmonit.com/monit/documentation/monit.html#program_status_testing