Re: Analysis of failing Nagios ping check in Wheezy

[Petter Reinholdtsen <pere@hungry.com>]

[Holger Levsen]
> so sitesummary could stay that way while sitesummary as we use it in
> debian- edu could only check via ipv4 ?! ;)

Sure.

But it will only hide one symptom of the real problem, which is that
mapping the local hostname to IP addresses return a invalid IPv6
address.  This can cause IPv6 enabled programs to fail unexpectedly.

This is how a sensible machine would behave:

  % getent hosts $(hostname)
  ::1             minerva.intern
  % ping6 -c3 $(hostname)
  PING minerva.intern(ip6-localhost) 56 data bytes
  64 bytes from ip6-localhost: icmp_seq=1 ttl=64 time=0.031 ms
  64 bytes from ip6-localhost: icmp_seq=2 ttl=64 time=0.058 ms
  64 bytes from ip6-localhost: icmp_seq=3 ttl=64 time=0.045 ms

  --- minerva.intern ping statistics ---
  3 packets transmitted, 3 received, 0% packet loss, time 1998ms
  rtt min/avg/max/mdev = 0.031/0.044/0.058/0.013 ms
  %

And this is how Debian Edu machines behave:

  root@tjener:~# getent hosts $(hostname)
  fe80::5652:ff:fe1f:e659 tjener.intern
  root@tjener:~# ping6 -c3 $(hostname)
  connect: Invalid argument
  root@tjener:~#

The fact is that 'check_ping -H $(hostname)' look up the name and end
up with a IPv6 address causes it to call ping6.  This will probably be
done by other programs too.

The real problem affect all IPv6 enabled programs.  Does it make sense
to only hide the symptom instead of fixing the cause?

-- 
Happy hacking
Petter Reinholdtsen