Strange ping problems

[gwakem]

RHEL 7.5, XI 5.6.2

Several days ago, I had two host checks go critical, when the hosts were up. I discovered that the IP of the host in the system contained a prepended 0 in the third octet, like so: 10.100.02.100. This had worked for several years up until this point. Even more strange, this is what the host status detail page says:

Code: Select all

check_ping: Invalid hostname/address - 10.100.09.100
Usage:
check_ping -H -w ,% -c ,%
[-p packets] [-t timeout] [-4

I ran the "Ping this host" and it returned "Ping output" and a blank box below. No error, and a link that says Ping another host. I removed the 0 from the third octet and it worked. I changed it and moved on.

Today the same thing happened, but I was able to find a host that still worked in this configuration.
Address: 10.100.07.100 - Host up. I ran the Ping this host:

Code: Select all

PING 10.100.07.100 (10.100.7.100) 56(84) bytes of data.
64 bytes from 10.100.7.100: icmp_seq=1 ttl=58 time=46.5 ms
64 bytes from 10.100.7.100: icmp_seq=2 ttl=58 time=32.1 ms
64 bytes from 10.100.7.100: icmp_seq=3 ttl=58 time=32.1 ms
64 bytes from 10.100.7.100: icmp_seq=4 ttl=58 time=32.1 ms
64 bytes from 10.100.7.100: icmp_seq=5 ttl=58 time=74.0 ms

--- 10.100.07.100 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4005ms
rtt min/avg/max/mdev = 32.137/43.419/74.078/16.315 ms

There is no apply or changes before the check goes critical. I am unsure as to what is happening here, or even where to start looking. Any advice?

[gwakem]

In the time it took to post this, four more popped up, with one showing this output (which is different from the rest)

Code: Select all

/bin/ping -n -U -W 30 -c 5 10.101.09.100
CRITICAL - Could not interpret output from ping command

[scottwilkerson]

I did some tests and on my newer CenOS 7.6 machine I am getting the same output as you when I have a leading 0 in one of the octets.

However on an older machine (CentOS 6) it was working fine.

It has to be a change is a system package as check_ping hasn't been updated in a really long time.

I'm trying to fund the culprit but it is hard to pinpoint

[scottwilkerson]

Another machine in-house we could reproduce it on had the following kernel

Code: Select all

[root@TEST_XI_CentOS_7 ~]# uname -a
Linux TEST_XI_CentOS_7 3.10.0-957.12.1.el7.x86_64 #1 SMP Mon Apr 29 14:59:59 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

Even more bizarre, I spun up a new VM and upgraded everything and it works as expect.

This machine has the following kernel and works fine

Code: Select all

uname -a
Linux localhost.localdomain 3.10.0-957.21.3.el7.x86_64 #1 SMP Tue Jun 18 16:35:19 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

Also on our machines where we reproduced it we get this

Code: Select all

# ping 10.100.09.100
ping: 10.100.09.100: Name or service not known

so, it isn't a Nagios problem but a problem with the OS determining that 10.100.09.100 is an IP address

[gwakem]

Ahhhh very interesting. That makes sense, thanks for investigating. This can be locked up.

[scottwilkerson]

Ahhhh very interesting. That makes sense, thanks for investigating. This can be locked up.

Locking