OSx86project IPv6-enabled

[Iljitsch van Beijnum]

Congratulations on becoming IPv6-enabled!

 

However, it doesn't work all that well for me... I can't seem to reach you over IPv6 using my normal setup. This is an IPv6 traceroute:

 

# traceroute6 -l www.osx86project.org

traceroute6 to www.osx86project.org (2002:42eb:b46b::2) from 2001:1af8:2:5::2, 64 hops max, 12 byte packets

1 2001:1af8:2:5::1 (2001:1af8:2:5::1) 0.586 ms 0.359 ms 0.355 ms

2 2001:1af8:1:15:20e:39ff:fe3c:2aaa (2001:1af8:1:15:20e:39ff:fe3c:2aaa) 1.105 ms 0.852 ms 0.850 ms

3 eth0-0-7.6b2.ams7.alter.net (2001:600:8:6::1) 1.233 ms 1.306 ms 1.230 ms

4 telecity.ams-ix.ipv6.network.bit.nl (2001:7f8:1::a501:2859:2) 1.983 ms 1.957 ms 1.981 ms

5 jun1.kelvin.ipv6.network.bit.nl (2001:7b8::290:6900:1cc6:d800) 102.430 ms 18.416 ms 3.226 ms

6 nlede01.sixxs.net (2001:7b8:3:4f:202:b3ff:fe46:bec) 3.235 ms 3.448 ms 3.228 ms

7 * * *

8 *^C

 

Looking at your address, you guys are getting your IPv6 connectivity from the "6to4" mechanism. When I use 6to4 too it works, but when I use "regular" IPv6 connectivity it doesn't. Very likely, there is a problem with the 6to4 gateway at your end. You can see if this is the case by trying a traceroute6 to a site with a regular IPv6 address (that doesn't start with "2002:"), such as www.runningipv6.net.

 

(When I tried to register for the forum or post this message over IPv6 that didn't work: "network connection lost".)

[Swad]

Thanks, Iljitsch. We'll look into it.

 

Others have reported HUGE speed gains when using IPv6. Here's what I wrote for the Front Page:

 

So by this point I’m sure you’ve realized that OSx86 Project is about one thing – providing technology news and discussion. In striving toward this goal, we thought we’d let you know that osx86project.org now offers IPv6 support for our visitors connected to Internet2. Though most IPv6 support is sporadic at best, it’ll soon propagate across the rest of the Net giving everyone a better browsing experience.

 

If you've got an IPv6-enabled connection, you're probably already enjoying our new setup. If not, check out our IPv4 portal that'll bump the site back down to a IPv4 setup. The frontend is identical, but new things are happening underneath.

 

We’re not aware of many other technology sites, let alone Mac sites, that have made the transition, although some may have done it in the background. The reason we’re letting you know about our work is to show that we’re committed to this kind of accessibility. We hope that other Mac and tech enthusiast sites are also making the change and ensuring that their sites are prepped for the future. Something like this may not be a huge leap forward in Internet technology, but these standards are only as good as the community that implements them.

 

Here’s hoping a small step like this will just make that community all the stronger.

[sHARD>>]

Congratulations on becoming IPv6-enabled!

 

However, it doesn't work all that well for me... I can't seem to reach you over IPv6 using my normal setup. This is an IPv6 traceroute:

 

# traceroute6 -l www.osx86project.org

traceroute6 to www.osx86project.org (2002:42eb:b46b::2) from 2001:1af8:2:5::2, 64 hops max, 12 byte packets

1 2001:1af8:2:5::1 (2001:1af8:2:5::1) 0.586 ms 0.359 ms 0.355 ms

2 2001:1af8:1:15:20e:39ff:fe3c:2aaa (2001:1af8:1:15:20e:39ff:fe3c:2aaa) 1.105 ms 0.852 ms 0.850 ms

3 eth0-0-7.6b2.ams7.alter.net (2001:600:8:6::1) 1.233 ms 1.306 ms 1.230 ms

4 telecity.ams-ix.ipv6.network.bit.nl (2001:7f8:1::a501:2859:2) 1.983 ms 1.957 ms 1.981 ms

5 jun1.kelvin.ipv6.network.bit.nl (2001:7b8::290:6900:1cc6:d800) 102.430 ms 18.416 ms 3.226 ms

6 nlede01.sixxs.net (2001:7b8:3:4f:202:b3ff:fe46:bec) 3.235 ms 3.448 ms 3.228 ms

7 * * *

8 *^C

 

Looking at your address, you guys are getting your IPv6 connectivity from the "6to4" mechanism. When I use 6to4 too it works, but when I use "regular" IPv6 connectivity it doesn't. Very likely, there is a problem with the 6to4 gateway at your end. You can see if this is the case by trying a traceroute6 to a site with a regular IPv6 address (that doesn't start with "2002:"), such as www.runningipv6.net.

 

(When I tried to register for the forum or post this message over IPv6 that didn't work: "network connection lost".)

 

We are indeed using 6to4 as our datacenter doesn't natively support IPv6 routing yet, so we must do this outside of our internal network. I've just tried a traceroute on my IPv6 setup (native, not tunneled), and it seems to be working. It could be a problem with sixxs routing. My traceroute runs directly through ucaid.edu Internet2, then hops on voxrox to the 6to4 link. Anyway, it's certainly odd it isn't working for you, I'll take a look.

[Iljitsch van Beijnum]

We are indeed using 6to4 as our datacenter doesn't natively support IPv6 routing yet, so we must do this outside of our internal network. I've just tried a traceroute on my IPv6 setup (native, not tunneled), and it seems to be working. It could be a problem with sixxs routing. My traceroute runs directly through ucaid.edu Internet2, then hops on voxrox to the 6to4 link. Anyway, it's certainly odd it isn't working for you, I'll take a look.

Recently, there have been reports of 6to4 troubles with several large networks in the US. I'm not entirely sure what's going on, though. But in general you're better off with a fixed or manual tunnel that you can get from a tunnel broker (see http://en.wikipedia.org/wiki/Tunnel_Broker ).

 

If you try to traceroute6 to me at 2001:1af8:2:5::2, how far do you get?

[sHARD>>]

Recently, there have been reports of 6to4 troubles with several large networks in the US. I'm not entirely sure what's going on, though. But in general you're better off with a fixed or manual tunnel that you can get from a tunnel broker (see http://en.wikipedia.org/wiki/Tunnel_Broker ).

 

If you try to traceroute6 to me at 2001:1af8:2:5::2, how far do you get?

 

I seem to have gotten all the way, but it took an impressive 20 hops. I'm talking with our 6to4 provider about the problem you're experiencing.

 

EDIT: It appears the sixxs tunnel has been having issues with their own, they've contacted sixxs in an attempt to get a fix. If you can route through anywhere else it should work fine, otherwise you're going to have to revert to IPv4, unfortunately. Our provider hopes to remove the 6to4 and go full IPv6 soon, hopefully reducing these routing issues.

[Fuzzel]

I seem to have gotten all the way, but it took an impressive 20 hops. I'm talking with our 6to4 provider about the problem you're experiencing.

 

EDIT: It appears the sixxs tunnel has been having issues with their own, they've contacted sixxs in an attempt to get a fix. If you can route through anywhere else it should work fine, otherwise you're going to have to revert to IPv4, unfortunately. Our provider hopes to remove the 6to4 and go full IPv6 soon, hopefully reducing these routing issues.

 

Who has contacted SixXS and what kind of "issues of their own" are they supposed to be having?

It is only that because I read NANOG that I noticed this issue, but there certainly has been no contacting towards SixXS about any issue. Next to that in the above traceroute SixXS doesn't even appear, except for the <a href="http://www.sixxs.net/faq/connectivity/?faq=6to4">6to4 gateway which coincidentially runs on the same machine as the nlede01 PoP</a>. SixXS itself doesn't do routing and only handles traffic of to the endtunnels.

Anyway claiming that you contacted SixXS is simply not true. Claiming that there are issues is also false. If there are then post those issues and contact the people who you think should be resolving them so that they can be resolved or the correct people can be contacted.

 

Now to solve your problem, instead of only replying on how lowly you are trying to be:

 

The problem you are having is the disfunctional <a href="http://www.6gate.com"></a> "IPv6 Gateway":

 

www.osx86project.org has AAAA address 2002:42eb:b46b::2

forum.osx86project.org has AAAA address 2002:42eb:b46b::2

 

This means that the website actually gets reached like:

Native IPv6 client -> ISP -> 6to4 gateway -> IPv6 in IPv4 (6to4) -> 6gate.com box [proxy] -> IPv4 -> Original Site

 

The problem can thus occur at many places. The traceroute which iljitsch posts stops at nlede01.sixxs.net, which

is the BIT 6to4 Gateway (and a SixXS PoP).

 

Thus lets see: IPv4 connectivity from that machine to the 6gate machine seems to work:

 

traceroute to 66.235.180.107 (66.235.180.107), 64 hops max, 40 byte packets

1 sixxs-gw.network.bit.nl (193.109.122.241) 0.579 ms 0.454 ms 0.353 ms

2 jun1.sara.network.bit.nl (213.136.31.3) 1.613 ms 10.435 ms 1.535 ms

3 106.ge4-0.mpr1.ams1.nl.above.net (62.93.194.34) 2.239 ms 2.078 ms 2.211 ms

4 so-3-0-0.cr1.lga1.us.above.net (64.125.27.185) 88.094 ms 88.750 ms 87.969 ms

5 so-1-0-0.mpr2.lga1.us.above.net (64.125.27.146) 88.307 ms 88.791 ms 88.127 ms

6 so-0-0-0.mpr2.lga3.us.above.net (64.125.30.22) 87.947 ms 88.153 ms 88.381 ms

7 so-0-0-0.mpr1.lga3.us.above.net (64.125.29.249) 87.720 ms 87.417 ms 87.767 ms

8 nyc1.ge7-0.core2.dca2.hopone.net (66.36.224.209) 89.421 ms 89.395 ms 89.800 ms

9 vl2.msfc1.distb1.dca2.hopone.net (66.36.224.228) 93.637 ms 92.613 ms 92.587 ms

10 project1.DNSstuff.com (66.235.180.107) 90.180 ms 90.148 ms 89.804 ms

 

But is apparently broken from various other places on the internet:

traceroute to 66.235.180.107 (66.235.180.107), 64 hops max, 40 byte packets

1 ge-1-3-0.colo.breda.concepts-ict.net (213.197.29.1) 0 ms 0 ms 0 ms

2 at-0-3-1.nikhef.concepts-ict.net (213.197.27.126) 2 ms 2 ms 2 ms

3 Gi0-0-203.ams-koo-access-2.interoute.net (212.23.59.161) 3 ms 3 ms 3 ms

4 PO6-0.ams-koo-core-2.interoute.net (212.23.41.145) [MPLS: Label 146 Exp 0] 79 ms 203

5 PO1-0.lon-wal-core-1.interoute.net (84.233.152.157) [MPLS: Label 229 Exp 0] 77 ms 77

6 PO5-0.lon-wal-core-2.interoute.net (84.233.152.162) [MPLS: Label 67 Exp 0] 77 ms 77

7 PO5-0.nyc-002-inter-1.interoute.net (212.23.43.133) [MPLS: Label 18 Exp 0] 77 ms 77

8 PO6-0.nyc-002-inter-2.interoute.net (212.23.43.130) 77 ms 77 ms 77 ms

9 * * *

10 *

 

traceroute to 66.235.180.107 (66.235.180.107), 30 hops max, 38 byte packets

1 swiCS3-G7-14.switch.ch (130.59.35.65) 0.236 ms 0.179 ms 0.164 ms

2 swiEZ2-P1.switch.ch (130.59.36.222) 0.255 ms 0.200 ms 0.200 ms

3 swiIX1-10GE-1-1.switch.ch (130.59.36.250) 0.290 ms 0.258 ms 0.256 ms

4 zch-b1-geth3-1.telia.net (213.248.79.189) 0.199 ms 0.187 ms 0.163 ms

5 ffm-bb1-link.telia.net (213.248.65.41) 5.921 ms 5.951 ms 5.918 ms

6 ffm-b3-pos0-1-0.telia.net (213.248.64.62) 6.323 ms 9.118 ms 42.911 ms

7 france-telecom-02189-ffm-b3.telia.net (213.248.77.206) 6.339 ms 9.100 ms 6.241 ms

8 *

9 *

10 *

 

Thus it gives to wonder how well the return path really is. But at least the nlede01.sixxs.net machine should be able to send packets towards the 6gate machine, lets see:

 

15:51:46.329838 IP 192.88.99.1 > project1.DNSstuff.com: IP6 2002:c058:6301::.53480 > 2002:42eb:b46b::2.33435: UDP, length 12

 

But never ever no return. Not even a port unreach. So it might quite well be that packets arrive at the destination but never find their way back at all.

 

And where it gets lost I will let to the magic fairys who think that 6to4 is easy to debug. I assume that somebody is wrongfully filtering out the 192.88.99.1 address...

As an easy fix 2002:42eb:a000::/35 is now routed as icmp unreach, thus giving browsers a chance to use IPv4.

 

Just in case you still do want to send a proper 'we have a message', see http://www.sixxs.net/contact/ it is really not that hard to find. Though for the 6to4 gateway you should be contacting noc@bit.nl.

[sHARD>>]

Anyway claiming that you contacted SixXS is simply not true. Claiming that there are issues is also false. If there are then post those issues and contact the people who you think should be resolving them so that they can be resolved or the correct people can be contacted.

 

I'm not going to respond to the rest of this junk, but I will say this:

 

1) How do you know who we have or haven't contacted?

2) The error occurs at sixxs, they have been contacted in order to attempt to debug the problem. No one said it was anyones fault

3) IPv6 was disabled awhile back in order for us to resolve these issues.

 

Please don't make false accusations, while you have a good grasp on IPv6, these simply make everyone look bad.

[Fuzzel]

I'm not going to respond to the rest of this junk, but I will say this:

 

1) How do you know who we have or haven't contacted?

Small hint: compare my name with <a href="http://www.sixxs.net/about/">SixXS about</a>.

 

2) The error occurs at sixxs, they have been contacted in order to attempt to debug the problem. No one said it was anyones fault

 

As explained it can't occur at SixXS as we don't operate any 6to4 relay. BIT does. See the above text.

 

3) IPv6 was disabled awhile back in order for us to resolve these issues.

 

Please don't make false accusations, while you have a good grasp on IPv6, these simply make everyone look bad.

 

I don't make any (false) accusations. The one accusing is the one saying that they contacted SixXS and that they are at fault, which they are not. You say again above "The error occurs at sixxs", now that is a false accusation.

 

As you say yourself, don't simply say that it is broken somewhere and make people look bad.

Read the post again and realize that I try to help you guys out. If I where accusing I would simply claim that it is all your fault, which it isn't One can't point at 'the fault' as 6to4 is something which one can't debug at all.

 

Use http://www.sixxs.net/contact/ if you need to contact SixXS, don't simply put out false accusations.

(one of the most important things in your false accusations btw is any debug information.... let alone that message that you supposedly sent to the SixXS staff, which you nor others really did not do...)

 

That said, there is nothing we can do about this issue. We might be able to do some debugging, but 6to4 is to weird to debug as it can go wrong in way too many places. As you don't have control of the 6gate.com servers there is no way to figure out what is going on on their end either, which makes it really difficult to debug all of this. Don't blame others if you don't know the cause, and especially if you didn't contact them at all....