host-check on multi interface server

[Rasmus Plewe]

On Mon, Feb 24, 2003 at 06:42:11PM +0100, Markus Schencker wrote:
> On Mon, Feb 24, 2003 at 15:37:55AM +0100, Rasmus Plewe wrote:
> 
> Example: server1 with the following interfaces:
> 172.19.0.1, 172.20.0.1, 172.21.0.1
> 
> define host server1 with IP address 172.19.0.1
> 
> define service ping1 on host server1 with IP address 172.19.0.1
> define service ping2 on host server1 with IP address 172.20.0.1
> define service ping3 on host server1 with IP address 172.21.0.1
> 
> define service disk1 on host server1 depends on ping1,ping2,ping3
> 
> If one of the ping tests fails the depending tests won't be executed.

Yes, but the disk1 could still be working and useful through other
interfaces. So, you basically drop dead if any of your interfaces die,
while I (see below) only drop dead if one specific interface dies. 
Which approach is more useful is, of course, open to discussion. ;-)

> > I defined the interfaces as hosts (with a check_ping as service) and
> > defined the server as parent. I believe this solution is fairly
> > common.
> 
> What is the advantage of your method? You define a logical grouping of the 
> interfaces. Which interface do you define as the parent server? Where do you 
> put your services (to the parent server host or to the interfaces)? What is 
> the result if one of the "interface" hosts will fail?

- Advantage: First idea I got when setting this up. Seemed logical to
  me and does exactly what I want, which is monitoring the whole
  system, but also generating only one notification even if the
  machine dies completely. 
- Parent server: doesn't really matter. The most important one or the
  most stable one. In any case, if only one interface dies I've got a
  problem (in my setup). They are not for redundancy. 
  In my case I put the parental host onto the administrative network,
  to minimize traffic for the production network. 
- Services: On the parent host. 
- Result of failing interfaces: I get a critical alert ("host down")
  and try to get the interface up again. As I'm a clever guy and know
  the network layout (or, rather, know where to look it up) I know
  what part of the system is affacted in which way, so when I can't
  simply get the interface up again I turn pale, start to reroute
  traffic and yell at the people in charge of replacing hardware... 


Rasmus


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
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