"funny" comment in pop-up display for a measurement? (v4/v6 addr mismatch)
Some more digging and some more confusion? Logged in, looking at "My measurements", selecting 1033944, a traceroutev6. Bringing up the Map tab, clicking on the green balloon in Bremen (Probe ID 13255) with the *2* (hops) in it, pop-up comes up. Nice! But the *-comment at the bottom puzzles me, by saying: * The last address does not match the target IP (192.153.174.196) This is sort of obvious, as the address referred to is 2001:6f8:114e:4::1 What's the background for this behaviour? Wilfried
Hi Wilfried, On 2014/01/15 18:44 , Wilfried Woeber wrote:
Some more digging and some more confusion?
Logged in, looking at "My measurements", selecting 1033944, a traceroutev6. Bringing up the Map tab, clicking on the green balloon in Bremen (Probe ID 13255) with the *2* (hops) in it, pop-up comes up. Nice!
But the *-comment at the bottom puzzles me, by saying:
* The last address does not match the target IP (192.153.174.196)
This is sort of obvious, as the address referred to is 2001:6f8:114e:4::1 What's the background for this behaviour?
The target of your measurement, wsww2.cc.univie.ac.at, has, as far as I can tell, only an IPv4 address. So an IPv6 measurement should fail. And in fact most probes reported: "result":[{"error":"name resolution failed: non-recoverable failure in name resolution"}] One question that comes up is how you were able to run this measurement in the first place. Because the system usually doesn't allow running measurement towards targets that don't have an address for the selected address family. This also why the UI complains about 192.153.174.196, because that is the IPv4 address of wsww2.cc.univie.ac.at. There were 3 probes where name resolution resulted in an IPv6 address. 2 of them have since changed their behavior and no longer resolve to an IPv6 address. One probe, 13255, is rather curious, because it resolved wsww2.cc.univie.ac.at to 2001:6f8:114e:3::c099:aec4, which is interesting because c099:aec4 is exactly equal to the IPv4 address of wsww2.cc.univie.ac.at. So I suspect that this probe is behind a resolver that does DNS64. Philip
Hi Philip thanks for the feedback! <blush> This was a very complex way to find out that I'm having a DNS (v6) config issue on my end. Sorry for the noise! Wilfried Philip Homburg wrote:
Hi Wilfried,
On 2014/01/15 18:44 , Wilfried Woeber wrote:
Some more digging and some more confusion?
Logged in, looking at "My measurements", selecting 1033944, a traceroutev6. Bringing up the Map tab, clicking on the green balloon in Bremen (Probe ID 13255) with the *2* (hops) in it, pop-up comes up. Nice!
But the *-comment at the bottom puzzles me, by saying:
* The last address does not match the target IP (192.153.174.196)
This is sort of obvious, as the address referred to is 2001:6f8:114e:4::1 What's the background for this behaviour?
The target of your measurement, wsww2.cc.univie.ac.at, has, as far as I can tell, only an IPv4 address. So an IPv6 measurement should fail. And in fact most probes reported:
"result":[{"error":"name resolution failed: non-recoverable failure in name resolution"}]
One question that comes up is how you were able to run this measurement in the first place. Because the system usually doesn't allow running measurement towards targets that don't have an address for the selected address family.
This also why the UI complains about 192.153.174.196, because that is the IPv4 address of wsww2.cc.univie.ac.at.
There were 3 probes where name resolution resulted in an IPv6 address. 2 of them have since changed their behavior and no longer resolve to an IPv6 address.
One probe, 13255, is rather curious, because it resolved wsww2.cc.univie.ac.at to 2001:6f8:114e:3::c099:aec4, which is interesting because c099:aec4 is exactly equal to the IPv4 address of wsww2.cc.univie.ac.at. So I suspect that this probe is behind a resolver that does DNS64.
Philip
Hi Wilfried, On 16.01.2014 13:18, Wilfried Woeber wrote:
[...]
This was a very complex way to find out that I'm having a DNS (v6) config issue on my end.
isn't debugging one's own end also an ATLAS goal? ;-) Best, -C.
On 2014/01/16 13:18 , Wilfried Woeber wrote:
This was a very complex way to find out that I'm having a DNS (v6) config issue on my end.
Well, the measurement was supposed to tell you that right away. But somehow we forgot to actually check for IPv6 addresses in this variation ('IPv6 Connectivity Test'). Philip
participants (3)
-
Carsten Schiefner -
Philip Homburg -
Wilfried Woeber