Nagios bizare ping behaviour

Question

I have nagios setup as I want it, but last night it started alerting me of a problem pinging the internet. (I have a check that pings www.google.com to test that the internet is still working)

Now looking into this problem I can see that my network can ping the internet fine, the server that runs nagios can also ping the internet but if I run

./check_ping -H www.google.com -w -c

I get Network Unreachable

but

./check_ping -H [IP address of google] -w -c

I get ping OK

I can use ping on both ip and domain name and they both work fine.

Anyone got any clues as to where the problem lies, is it nagios or do I have a problem with DNS in my nagios box or in my network somewhere?

Thanks

Solution 1

Same thing happened to us last night starting at about 20:01 EDT, and the alert is still not functioning properly. I hadn't tried running the check_ping with one of Google's IP addresses until I read your post though, and it worked here just as it did for you. ping www.google.com returns successful pings, and I can browse to www.google.com from my nagios server as well, but the check_ping monitor is still critical.

-- fixed-for-me --

I found that the ping6 www.google.com command failed immediately with a network unreachable error, while a ping www.google.com succeeded. I ran the check_ping command as before with www.google.com adding a -4 flag at the end to force it to use IPV4, and the check_ping now succeeds.

/usr/lib/nagios/plugins/check_ping -H www.google.com -w 100.0,20% -c 200.0,40% -4
PING OK - Packet loss = 0%, RTA = 22.51 ms

So I edited the commands.cfg file to add a -4 to the end of the command, and voila, a successful check.

Solution 2

Glad you're participating in the IPv6 day and the IPv6 launch of the "most important internet sites". You are probably pinging a v6 site address when using the name without having v6 connectivity at your Nagios server. The good news is that it is over by midnight :)

Solution 3

The best way to solve this problem is to use

check_command                   check-host-alive_4

host

check-host-alive_4 is defined in /etc/nagios-plugins/config/ping.cfg (in Ubuntu 14.04) specially for occasions described by Shawn Maceno above.

Comments

[wolfgangsz]

What happens when you do "host www.google.com"?

[Simon Foster]

Glad to here I am not alone, which suggests that it might not be my network this time that is the problem

[Hank]

On debian, the nagios-plugins-basic package (provder of the check-host-alive definition) also defines check-host-alive_4.