Nagios Core 4.1.1 - false check_ping, check_icmp results

[gmarks]

As I said earlier, the packets are not leaving Nagios host's interface/NIC, they are not going out on the network so the problem is definitely on the source.
Instead of 30 ICMP requests, tcpdump catches less, and at that occasions the results calculations are incorrect.
There are no drops/errors on that interface.
Using check_icmp only caused incorrect results for ICMP based checks, so changing the protocol doesn't seem like the right direction.
When check_icmp is not used anywhere, only check_ping, everything works well.
Thanks once again for your help, I appreciate it.

[eloyd]

As I said earlier, the packets are not leaving Nagios host's interface/NIC, they are not going out on the network so the problem is definitely on the source.

I'm sure you're right, but how are you confirming this? Are you checking on the Nagios server that the packets are being generated? As in, tcpdump dst X.X.X.X (use the IP of your destination host)?

[gmarks]

This is the command I used :

Code: Select all

tcpdump '((icmp) and ((src host 10.60.60.228) or (dst host 10.60.60.228)))' -w icmp.pcap

Where 10.60.60.228 is the IP of my Nagios host.

[rkennedy]

Got it. Thanks for describing this with such detail.

Can you open a bug report on https://github.com/nagios-plugins/ for this seperate issue, or would you like me to?

[gmarks]

If it's not neccessary for someone else to confirm to have same issues, I would appreciate if you'd open the bugreport.
Thanks.

[rkennedy]

I'm having a hard time replicating this. I have no problem filing the bug report, but I'll need your assistance.

- I have a command defined (check_icmp2) that uses /usr/local/nagios/libexec/check_icmp running against 192.168.3.243
- From 192.168.4.254 I have it using check_icmp2, forcing a check when it's up results in OK - 192.168.3.243: rta 0.982ms, lost 0%
- Forcing a check when it's down, results in CRITICAL - 192.168.3.243: rta nan, lost 100%

From what I understand, in your case the down host is not reporting 100% lost. Is that right? Could you point me in the right direction to replicating this on what I'm not doing right?