clarification on parents and dependencies

Matt Pounsett matt.pounsett at cira.ca
Tue Apr 13 17:59:14 CEST 2004


On Mon, 12 Apr 2004, Ivan Fetch wrote:

> > Nagios -> Switch A -> Switch B -> Sun Workstation
> >
>    Yes, exactly.
> 
> 
> > If it's the first one, then Switch B is the parent of the Workstation, and
> > Switch A is the parent of Switch B.
> 
>    Ok, will switch a need to be the parent of all other switches
> because the Nagios box is connected to switch a, even if both switch a and
> switch b are connected by fiber to the router (vs. fiber connecting the
> switches directly)?

I'm not sure I follow this.  In my diagram above, the workstation would not
list switch A as a parent, only switch B.  Switch A is, by implication, a
grandparent of the workstation.   If switch A were to go down, both switch B
and the workstation would be marked as "unreachable" by Nagios.

> > >    2. Our switches are in a different sub net (apparently in order to be
> > > trunked), so if something is going wrong with our router, we can not ping
> > > the switches.  Would folks out there recommend specifying the router as a
> > > parent to
> > > each switch's host definition?  I see a kind of loop potential in that
> >
> > This one is more compex.. it sounds like you've got some VLANs on the switch,
> > and the packets pass through the switch a couple of times on their way to
> > their destination, right?
> 
> Yes, the switches are trunked and getting their vlans from the router, so
> even for local trafic on one switch to get where it's going, it looks like
> communication with the router is necessary.  IF the router is experiencing
> high load or some other problem, local trafic on a switch is still
> effected (although none of that traffic should have to go anywhere except
> from one fast ethernet port to another).  This makes it difficult for
> Nagios to know whether the problem is the switch, the router, or a
> host/service.  Since I can not ping the switch in this kind of a situation
> (because the switch is in another subnet and has to go through the router)
> perhaps incorparating the switches into my monitoring will not be
> possible, or will not be very reliable?

If problems with the router will cause the switch to be unreachable, then the
router should be the switch's parent.  Remember, the parent/child
relationships are more about dependancy than they are about the network
physical or logical layout.  You have a bit of a unique situation where your
switch may be rendered useless by the router, but the router is only reachable
through the switch... which means they are really co-dependant, and Nagios
doesn't have a way to diagram that.  I'd say it's a judgement call which
device you make the parent.   Just make sure that anyone who may be
potentially paged about those two devices know about the configuration, and
that it might be either device which is affected.

-- 
Matt Pounsett                 CIRA - Canadian Internet Registration Authority
Technical Support Programmer                    350 Sparks Street, Suite 1110
matt.pounsett at cira.ca                                 Ottawa, Ontario, Canada
613.237.5335 ext. 231                                      http://www.cira.ca



-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null





More information about the Users mailing list