pcp
[Top] [All Lists]

Wierd pmie bug of the day?

To: kmcdonell@xxxxxxxxxx
Subject: Wierd pmie bug of the day?
From: Nathan Scott <nscott@xxxxxxxxxx>
Date: Wed, 11 Apr 2007 16:25:18 +1000
Cc: pcp@xxxxxxxxxxx
Organization: Aconex
Reply-to: nscott@xxxxxxxxxx
Sender: pcp-bounce@xxxxxxxxxxx
Hi Ken,

I'm struggling to make pmie honour some of its time-based macros;
in particular the following rule pointed at an archive seems to be
constantly getting a value of zero for $hour, constantly...

// Host CPU usage during office hours on work days
(((kernel.all.cpu.user + kernel.all.cpu.intr +
   kernel.all.cpu.nice + kernel.all.cpu.sys) / hinv.ncpu) > 0)
 && ($day_of_week > 0 && $day_of_week < 6)
 && ($hour == 0)
// && ($hour >= 09 && $hour <= 17)
-> print "%v\n";

The expression I want is the commented out one, but the only
value of $hour that seems to give truth is 0 ... same for $minute.
Am I using these keywords incorrectly or is this a pmie buglet?

>From poking around in gdb a bit, reflectTime does seem to be called,
and does seem to update the global minute/hour doubles correctly,
but somewhere in the expression tree evaluation below eval(), the
wheels seem to be falling off, maybe.

thanks!

-- 
Nathan


<Prev in Thread] Current Thread [Next in Thread>
  • Wierd pmie bug of the day?, Nathan Scott <=