On Mon, Jan 30, 2012 at 11:39 AM, Claudio Kuenzler <span dir="ltr"><<a href="mailto:ck@claudiokuenzler.com">ck@claudiokuenzler.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
It looks like you use parent-child-relationships?<br>That would explain why it's written "Host Unreachable (172.21.10.170) " instead of Host Down. <br>Was the parent of this host down as well? In this case, Nagios 'knows' that it doesn't make sense to continue checks on the child hosts.<br>
</blockquote><div><br></div><div>This host is duel homed using both Ethernet and Infiniband networks. To Nagios's point of view the parent of this host is the Ethernet switch that it is plugged into, which didn't go down at all during this outage. Host Unreachable is a valid ICMP code, so that's why the plugin reported this error, not because of any Nagios parent/child relationship.</div>
<div><br></div><div>So in this situation Nagios should have increased the check attempts count by one until the max check attempts threshold was reached, thus sending out a service notification. Is this an error in Nagios or the check_ping plugin. I'm using Nagios v3.2.3 and check_ping v1.4.15 by the way.</div>
<div><br></div><div>--<div><b><font face="arial, helvetica, sans-serif">Robert V. Bolton</font></b></div><div><font face="arial, helvetica, sans-serif">Email: <a href="mailto:robert@robertvbolton.com" target="_blank">robert@robertvbolton.com</a></font></div>
<div>Web: <a href="http://robertvbolton.com" target="_blank">http://robertvbolton.com</a></div></div></div>