[Nagios-devel] Initial Scheduling Problem explanation

[Russell S. <ru...@qu...>]

Hey Darren, I got this as a response directed to me instead of the list. 
 I sent it out a while ago, but you might have mised it.  This might 
help you understand the behavior you're getting.

-Russell Scibetti

-------- Original Message --------
Subject: Re: [Nagios-devel] Re:initial scheduling problem
Date: Wed, 17 Jul 2002 16:40:11 -0500
From: Bill Huff <bh...@co...>
To: Russell Scibetti <ru...@qu...>
References: <078EC26E265CD411BD9100508BDFFC861BCBB3F0@shawmail02> 
<3D3...@qu...>



  Russell, I am surprised that someone hasn't answered this by now,
  so let me give it a shot.  With a service_interleave_factor of
  's', that means that Nagios will us its smart method to schedule
  'initial' service checks.  How it does this is pretty well
  documented here:

    http://nagios.sourceforge.net/docs/1_0/checkscheduling.html#inter_check_delay

  I will give a shot at making it short and sweet.  For initial
  scheduling nagios will add the check interval for all services and
  come up with a 'total check time'.  It then divides that number
  by the total number of services to be checked to come up with
  the inter_check_delay.  Once it determines the average check time
  it then figures out how many checks it has to do a second to test
  all of them in that average time.  This all works great, if your
  services all have similar check intervals.  However, if you do
  something like have one service that checks once every 2 hours,
  it totally blows the average and will cause Nagios to space out
  the initial service check schedule over a much wider time then
  it should.  Do you have any checks that are being tested much
  less frequently then the rest?  That could be causing what you
  are seeing.

  I ran into this problem when I had a check that was going
  once every 6 hours and the rest were all going once every 1 or
  2 minutes.  It caused then initial checks to be spaced out over
  about 30 minutes instead of the 2 that I would have expected.

  However, if you have 'retain_state_information' set to a 1 and
  you simply add 1 new service and the same thing happens, then I
  would say that that would be a bug, since Nagios shouldn't have
  to schedule all of the services initial checks, only the new one,
  so it should be almost immediate.

  Hopefully this will be helpful.  Take a look at the documentation
  that I reference above and if you have any other questions feel
  free to ask and I will try and make myself more clear.

--
Bill