Host vs Service reachability

[Jon Angliss]

On Fri, 15 May 2009 11:26:54 -0400, Jason Frisvold
<frisvolj at lafayette.edu> wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>Hi all,
>
>	I'm looking at possibly disabling my service pings and moving over
>completely to host checks for reachability.  I'm unsure of the exact
>mechanics, though, so if someone can help me out, I'd appreciate it.
>
>	So, to start, if a host has no current services, is the host_check
>still intermittent?  Or does it become an active check at a specific
>interval, similar to a service?  We have several hosts that we only
>check for connectivity as opposed to services, for a variety of reasons.

>From my understanding, the changes in v3 make the checks perform
better, whilst v2 were done slightly differently.  They execute as
scheduled based on the check interval.

>	From what I've read of Nagios 3, it sounds like using host checks is
>more efficient.  Are there instances, though, where services may show
>up, but the host is down?  Does the host_check run at specified
>intervals, even though service checks don't show a problem?

A service may show up, and a host down if the times for checking
overlap.  For example, a service check every 10 minutes, and host
checks every 1 minute.  A host check could show the server down,
whilst the service check will show it is OK.

It can also depend on what you consider for your host check too.  For
example, if you cannot ping, and you use the check_ping/check_icmp
functions, you'll show the host down, but services up.

>	Is there a compelling reason to not do this?  ie, am I inviting a world
>of hurt by using host checks vs a service ping?

You're actually "inviting a world of hurt" doing it the other way
around (service ping over host check).  Host checks will help surpress
service notifications, as well as help manage network reachability if
you use parents (see [1]).  Take for example, a host you have FTP,
HTTP, and your ping check on.  If you the host goes down, the ping
service will show down, but FTP, and HTTP tests will also show
unreachable, or timeout, causing 3 notifications.  If the host check
was to be used, you'd get a notification the host was down, rather
than 3 to report host issues.  You can, of course, work around this
with service dependencies, but that complicates your configuration.



[1]: http://nagios.sourceforge.net/docs/3_0/networkreachability.html

-- 
Jonathan Angliss
<jon at netdork.net>


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables 
unlimited royalty-free distribution of the report engine 
for externally facing server and web deployment. 
http://p.sf.net/sfu/businessobjects
_______________________________________________
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