OpenBSD Journal

IPv6 Connectivity for Undeadly

Contributed by merdely on from the wwwv6 dept.

IPv6 connectivity is now enabled for the OpenBSD Journal, try http://ipv6.undeadly.org/ (which resolves to 2001:4830:155e::1). Since it's being tunneled through IPv4, it's probably slower than IPv4.

Certain features such as logging of IPv6 addresses in comments, https authentication, etc., could still use some testing. If you have IPv6 connectivity, please give http://ipv6.undeadly.org/ a run through of the features you normally use and leave feedback below.

Also, leave a comment on why you are using IPv6 and whether you think it still is the next great thing or if it has failed.

(Comments are closed)


Comments
  1. Comments
    1. By Janne Johansson (2001:6b0:5:1095:20d:56ff:fe11:3120) jj@inet6.se on www.inet6.se

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      Seems to work.

    2. By Anonymous Coward (2001:6f8:900:896::2) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/
      test test :-p

    3. By Anonymous Coward (2001:6f8:927:0:1163:5c4c:9d89:5cec) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      Looks good!

    4. By jirib (2001:15c0:65ff:ff::2) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      elinks & openbsd 4.1 :)

    5. By Anonymous Coward (69.71.33.160) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      Looking good.

    6. By Reza Muhammad (2001:328:2002:f107:211:2fff:fe39:51ae) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      It took me a while to update the AAAA record for the hostname, but it looks fine now :D

    7. By Todd T. Fries (2001:240:58a::3f) todd@fries.net on http://todd.fries.net/

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      It works with one caveat. 'host -C undeadly.org' shows dns servers are not in sync, so the new IPv6 records are not resolving everywhere yet.
      Ohwell, I put the IPv6 address in /etc/hosts and verified that the virtualhost has undeadly.org as well as ipv6.undeadly.org defined. Looking forward to undeadly.org having a v6 address ;-)

    8. By Brad (brad) brad at comstyle dot com on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      I am looking forward to (www.)undeadly.org having AAAA records.

      Comments
      1. By Daniel Hartmeier (2001:6f8:1098::3) daniel@benzedrine.cx on

        > I am looking forward to (www.)undeadly.org having AAAA records.

        Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?

        Comments
        1. By Todd T. Fries (2001:240:58a::1c) todd@fries.net on http://todd.fries.net/

          > From Brad:
          > > I am looking forward to (www.)undeadly.org having AAAA records.
          >
          From Daniel:
          > Given that IPv6 is inherently slower (due to the tunneling) still,
          > and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?

          If that was not a good idea, then getaddrinfo() and friends and BCP's would not suggest this behavior.

          The idea is, if someone has v6 or v4 or both, they have a list of addresses to connect to, and will try the v6 first. If they have no route to the v6 address, then they use v4.

          For the future transition to v6, if everybody (like silc in ports, hello!) tried IPv4 by default, we'd have to tweak software to no end just to tryout IPv6. Is this backwards or what?

          No, the v6 by default if you have a v6 route is best common practice.

          For whatever reason, if people have issues, for domains I host, I always put v6.domain.com with v6 only and v4.domain.com for v4 only, and (www.)domain.com with v6 and v4. This way, by default, things should just work, and if people for some reason _need_ to go to the site with one address family or another, they don't have to reconfigure their software to do it. For example:

          http://openbsd.fries.net (v4+v6)
          http://v6.openbsd.fries.net (v6 only)
          http://v4.openbsd.fries.net (v4 only)

        2. By Paul de Weerd (2001:980:fff:28:a00:20ff:feda:dde) weerd@weirdnet.nl on

          > > I am looking forward to (www.)undeadly.org having AAAA records.
          >
          > Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?

          It's actually the reason why firefox prefers IPv6. Yes, currently most IPv6 traffic is tunneled, but hopefully, one day, the situation will be reversed (for most of us, some already have v6-only links).

          Cool stuff ! Once I have a fixed v4 address at home, I should get myself another tunnel ;)

          Comments
          1. By Paul de Weerd (2001:a60:f044::) weerd@weirdnet.nl on

            > Cool stuff ! Once I have a fixed v4 address at home, I should get myself another tunnel ;)

            Until such a time, I double-tunnel from my dynamic IP ;)

            I'll take the added latency happily if it gives me v6 at home !

        3. By mk (130.225.243.71) on

          > > I am looking forward to (www.)undeadly.org having AAAA records.
          >
          > Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?

          Daniel, have you done any benchmarks to see the difference?

          Obviously there are at least two parameters, latency and throughput. When I had a tunnel a few years ago, I could not measure any difference in throughput. I never measured latency, but I didn't notice it being higher when I logged on to a v6-enabled machine a friend of mine runs.

          Just a thought. If the difference is negligible, setting AAAA records for undeadly.org would be nice.

        4. By Renaud Allard (2001:618:400::55c9:3f27) on

          > > I am looking forward to (www.)undeadly.org having AAAA records.
          >
          > Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?
          >
          >
          Is it that slower?

          $ ping6 ipv6.undeadly.org
          PING6(56=40+8+8 bytes) 2001:618:400::55c9:3f27 --> 2001:4830:155e::1
          16 bytes from 2001:4830:155e::1, icmp_seq=0 hlim=56 time=212.051 ms
          16 bytes from 2001:4830:155e::1, icmp_seq=1 hlim=56 time=209.103 ms

          $ ping www.undeadly.org
          PING www.undeadly.org (66.181.209.96): 56 data bytes
          64 bytes from 66.181.209.96: icmp_seq=0 ttl=242 time=174.863 ms
          64 bytes from 66.181.209.96: icmp_seq=1 ttl=242 time=168.117 ms

          Furthermore, most people using ipv6 nowadays know the weaknesses of running it.

          Comments
          1. By Daniel Hartmeier (193.192.245.3) daniel@benzedrine.cx on

            > Furthermore, most people using ipv6 nowadays know the weaknesses of running it.

            Alright, both undeadly.org and www.undeadly.org now have AAAA records :)

            Comments
            1. By Ryan McBride (2001:240:688:98::1) on


              > Alright, both undeadly.org and www.undeadly.org now have AAAA records :)

              Excellent, thanks Daniel. It's much better not to have to select special URLs to conduct v6 testing, it should Just Work(tm).

              Comments
              1. By Anonymous Coward (2001:6f8:94d:5::2) on

                It does Just Work(TM) now, for me.

          2. By Samhain (samhain) on

            > > > I am looking forward to (www.)undeadly.org having AAAA records.
            > >
            > > Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?
            > >
            > >
            > Is it that slower?
            >
            > $ ping6 ipv6.undeadly.org
            > PING6(56=40+8+8 bytes) 2001:618:400::55c9:3f27 --> 2001:4830:155e::1
            > 16 bytes from 2001:4830:155e::1, icmp_seq=0 hlim=56 time=212.051 ms
            > 16 bytes from 2001:4830:155e::1, icmp_seq=1 hlim=56 time=209.103 ms
            >
            > $ ping www.undeadly.org
            > PING www.undeadly.org (66.181.209.96): 56 data bytes
            > 64 bytes from 66.181.209.96: icmp_seq=0 ttl=242 time=174.863 ms
            > 64 bytes from 66.181.209.96: icmp_seq=1 ttl=242 time=168.117 ms
            >
            > Furthermore, most people using ipv6 nowadays know the weaknesses of running it.

            Well, that depends:

            bors# ping undeadly.org
            PING undeadly.org (66.181.209.96): 56 data bytes
            64 bytes from 66.181.209.96: icmp_seq=0 ttl=244 time=46.354 ms
            64 bytes from 66.181.209.96: icmp_seq=1 ttl=244 time=46.490 ms
            64 bytes from 66.181.209.96: icmp_seq=2 ttl=244 time=46.701 ms
            --- undeadly.org ping statistics ---
            3 packets transmitted, 3 packets received, 0.0% packet loss
            round-trip min/avg/max/std-dev = 46.354/46.515/46.701/0.142 ms
            bors# ping6 ipv6.undeadly.org
            PING6(56=40+8+8 bytes) 2001:470:1f03:f0a::2 --> 2001:4830:155e::1
            16 bytes from 2001:4830:155e::1, icmp_seq=0 hlim=57 time=267.947 ms
            16 bytes from 2001:4830:155e::1, icmp_seq=1 hlim=57 time=268.894 ms
            16 bytes from 2001:4830:155e::1, icmp_seq=2 hlim=57 time=268.373 ms

            Your IPv4 seems to be really slow.

            Comments
            1. By Steven Stillaway (samhain) on Steven Stillaway

              Hmm, it also did not seem to list my IP in the comment above.

              > > > > I am looking forward to (www.)undeadly.org having AAAA records.
              > > >
              > > > Given that IPv6 is inherently slower (due to the tunneling) still, and Mozilla prefers IPv6 over IPv4 automatically, is that a good idea?
              > > >
              > > >
              > > Is it that slower?
              > >
              > > $ ping6 ipv6.undeadly.org
              > > PING6(56=40+8+8 bytes) 2001:618:400::55c9:3f27 --> 2001:4830:155e::1
              > > 16 bytes from 2001:4830:155e::1, icmp_seq=0 hlim=56 time=212.051 ms
              > > 16 bytes from 2001:4830:155e::1, icmp_seq=1 hlim=56 time=209.103 ms
              > >
              > > $ ping www.undeadly.org
              > > PING www.undeadly.org (66.181.209.96): 56 data bytes
              > > 64 bytes from 66.181.209.96: icmp_seq=0 ttl=242 time=174.863 ms
              > > 64 bytes from 66.181.209.96: icmp_seq=1 ttl=242 time=168.117 ms
              > >
              > > Furthermore, most people using ipv6 nowadays know the weaknesses of running it.
              >
              > Well, that depends:
              >
              > bors# ping undeadly.org
              > PING undeadly.org (66.181.209.96): 56 data bytes
              > 64 bytes from 66.181.209.96: icmp_seq=0 ttl=244 time=46.354 ms
              > 64 bytes from 66.181.209.96: icmp_seq=1 ttl=244 time=46.490 ms
              > 64 bytes from 66.181.209.96: icmp_seq=2 ttl=244 time=46.701 ms
              > --- undeadly.org ping statistics ---
              > 3 packets transmitted, 3 packets received, 0.0% packet loss
              > round-trip min/avg/max/std-dev = 46.354/46.515/46.701/0.142 ms
              > bors# ping6 ipv6.undeadly.org
              > PING6(56=40+8+8 bytes) 2001:470:1f03:f0a::2 --> 2001:4830:155e::1
              > 16 bytes from 2001:4830:155e::1, icmp_seq=0 hlim=57 time=267.947 ms
              > 16 bytes from 2001:4830:155e::1, icmp_seq=1 hlim=57 time=268.894 ms
              > 16 bytes from 2001:4830:155e::1, icmp_seq=2 hlim=57 time=268.373 ms
              >
              > Your IPv4 seems to be really slow.
              >
              >

            2. By Renaud Allard (85.201.63.39) on

              > Your IPv4 seems to be really slow.
              >

              I am in Belgium. Also, your tunnel may be slower than mine.

    9. By Renaud Allard (2001:618:400::55c9:3f27) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      Looks like it works like a charm

    10. By Pat Thoyts (2001:960:79b:0:8d49:9ba1:e543:8049) on

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/
      Works for me.

    11. By Olof (2001:16d8:ff3c:201::2) olof@naya.se on www.naya.se

      > Testing IPv6 connectivity, http://ipv6.undeadly.org/

      Good work, always nice with some new sites IPv6 enabled.

  2. By Anonymous Coward (2001:4830:123a:beef:20e:a6ff:fe14:9a5b) on

    anybody got an URL of ipv6 patches for openbsd41 apache?

    Comments
    1. By Todd T. Fries (2001:240:58a::3f) todd@fries.net on http://todd.fries.net/

      > anybody got an URL of ipv6 patches for openbsd41 apache?

      This applies to current, I only guarantee it compiles. The README.v6 is the original README; php modules work fine in production with this, as does mod_ssl, but mod_perl does not load.

      http://FreeDaemonConsulting.com/drop/apachev6-4.2cur.diff

      P.S. Does anyone have a v6 diff for webalizer?

      Comments
      1. By Anonymous Coward (2001:4830:123a:beef:20e:a6ff:fe14:9a5b) on

        > anybody got an URL of ipv6 patches for openbsd41 apache?
        >
        > This applies to current, I only guarantee it compiles. The README.v6 is the original README; php modules work fine in production with this, as does mod_ssl, but mod_perl does not load.
        >
        > http://FreeDaemonConsulting.com/drop/apachev6-4.2cur.diff
        >
        > P.S. Does anyone have a v6 diff for webalizer?

        It applied and compiled onto 41-stable as well. mod_ssl works, will test php later on.
        Thanks.

      2. By Anonymous Coward (85.178.99.194) on

        > anybody got an URL of ipv6 patches for openbsd41 apache?
        >
        > This applies to current, I only guarantee it compiles. The README.v6 is the original README; php modules work fine in production with this, as does mod_ssl, but mod_perl does not load.
        >
        > http://FreeDaemonConsulting.com/drop/apachev6-4.2cur.diff
        >
        > P.S. Does anyone have a v6 diff for webalizer?

        Wouldn't be thttpd the best choice?
        Drop Apache, get thttpd? It even is BSD-Licensed :-D

        It's a pain in the ass to read CVS-Commits where Bugs get fixed wich where fixed in a "regular Apache" some time ago.
        To have a forked Apache is kinda useless because the code itself is "ugly" (like some developers said...) and thus leads to maybe less interest to improve it even.

        So I personaly would like to see a serious talk about the software in the base system and how/if things should change/get replaced for OpenBSD 4.3

        I still wonder why openhttpd.org was ever registered if it's not used at all.

  3. By almeida (almeida) on

    I have an IPv6 tunnel setup, but I think Undeadly is probably only the second IPv6 server I've hit in the whole time I've been running it.

  4. By Aaron Linville (2001:470:1f01:555:20d:93ff:fec2:d464) aaron@linville.org on http://www.linville.org/

    Cool, Undeadly is now the first site I visit regularly that has IPv6 connectivity! Now if we could just get Ars, Digg or /. into the 21st century.

    Comments
    1. By Todd T. Fries (2001:240:58a::3f) todd@fries.net on http://todd.fries.net/

      > Cool, Undeadly is now the first site I visit regularly that has IPv6 connectivity! Now if we could just get Ars, Digg or /. into the 21st century.

      If you choose, you can access any website via IPv6 by adding .sixxs.org to the url. See http://www.sixxs.net/misc/coolstuff/ for details. For example:

      http://arstechnica.com.sixxs.org/articles/paedia/IPv6.ars, http://digg.com.sixxs.org/linux_unix/IPv6_Connectivity_for_Undeadly, or http://slashdot.org.sixxs.org/articles/07/08/02/1237239.shtml.

      For bonus reading:

      http://www.potaroo.net.sixxs.org/tools/ipv4/index.html.

  5. By Anonymous Coward (2001:388:c01d:0:20e:cff:fe73:8ed9) on

    If only this were posted via native IPv6 the entire route, as opposed to having to traverse two tunnels (mine and undeadly's)

    This still deserves kudos though!

  6. By Anonymous Coward (24.22.214.92) on

    Unfortunately my ISP does not support this :(

  7. By jirib (195.212.29.163) on

    and www.openbsd.org will not become IPv6 ready? :)

  8. By Peter Curran (2001:4b10:100d:1:230:1bff:feb5:defb) peter@closeconsultants.com on

    Works seamlessly for me - I have native v6 from my ADSL provider.

    PING www.undeadly.org (66.181.209.96) 56(84) bytes of data.
    64 bytes from ip-66-181-209-96.tera-byte.com (66.181.209.96): icmp_seq=1 ttl=247 time=154 ms
    64 bytes from ip-66-181-209-96.tera-byte.com (66.181.209.96): icmp_seq=2 ttl=247 time=153 ms
    64 bytes from ip-66-181-209-96.tera-byte.com (66.181.209.96): icmp_seq=3 ttl=247 time=154 ms

    --- www.undeadly.org ping statistics ---
    3 packets transmitted, 3 received, 0% packet loss, time 2001ms
    rtt min/avg/max/mdev = 153.402/154.062/154.694/0.527 ms
    peter@dogmatix:~> ping6 www.undeadly.org
    PING www.undeadly.org(ipv6.undeadly.org) 56 data bytes
    64 bytes from ipv6.undeadly.org: icmp_seq=1 ttl=54 time=202 ms
    64 bytes from ipv6.undeadly.org: icmp_seq=2 ttl=54 time=200 ms
    64 bytes from ipv6.undeadly.org: icmp_seq=3 ttl=54 time=201 ms

    --- www.undeadly.org ping statistics ---
    4 packets transmitted, 3 received, 25% packet loss, time 2999ms
    rtt min/avg/max/mdev = 200.407/201.535/202.732/0.950 ms

    Bit of an overhead showing up - could be the path....

    Comments
    1. By Mark Shroyer (Niten) on http://markshroyer.com/

      > Works seamlessly for me - I have native v6 from my ADSL provider.

      Out of curiosity, who is your ISP? I wish my cable company would offer dual IPv4/v6 service...

      Comments
      1. By dingo (70.8.120.102) on

        > > Works seamlessly for me - I have native v6 from my ADSL provider.
        >
        > Out of curiosity, who is your ISP? I wish my cable company would offer dual IPv4/v6 service...

        i second that -- what country do you live in?

        Comments
        1. By Chmeee (85.136.52.161) on

          > > > Works seamlessly for me - I have native v6 from my ADSL provider.
          > >
          > > Out of curiosity, who is your ISP? I wish my cable company would offer dual IPv4/v6 service...
          >
          > i second that -- what country do you live in?
          Acording to 'whois 2001:4b10:100d:1:230:1bff:feb5:defb'

          inet6num: 2001:4b10::/32
          netname: UK-BOGONS-20041222
          descr: Bogons Ltd
          country: GB

          This leads us to http://www.bogons.net/
          "Internet Services for the Clueful"

          Chmeee

          Comments
          1. By sthen (85.158.44.149) on

            > > > > Works seamlessly for me - I have native v6 from my ADSL provider.

            How do you connect, pppoe(4), pppoe(8), something else?

            > This leads us to http://www.bogons.net/
            > "Internet Services for the Clueful"

            And damn good they are, too.

  9. By Michiel van Baak (mvanbaak) undeadly@vanbaak.info on http://michiel.vanbaak.info

    yay ! works great here.
    Thanks.

    I use ipv6 at home for my irc server and ftp server.
    All our professional hosted services are both ipv4 and ipv6 reachable.
    It's still hardly used, but we slowely see more and more hits on the ipv6 part of the services.

    Comments
    1. By Daniel Hartmeier (2001:6f8:1098::3) daniel@benzedrine.cx on

      > It's still hardly used, but we slowely see more and more hits on the ipv6 part of the services.

      I updated the pfstat graphs to keep track of the percentage of IPv6 traffic. So far, it looks like it's about 10% (IPv4 numbers include the encapsulated IPv6 traffic).

      I guess IPv6 adoption is higher than average among our readers ;)

  10. By Matthieu Herrb (2001:7a8:70b8:1:217:f2ff:fe41:7411) on

    Test 1 2 3

    I've IPv6 connectivity though my ISP and use it to test various software for v6 capability. Still handfuls of bugs to be found before it's available for general use, especially when dual stacks are involved.

    One thing that bothers me: when both AAAA and A records are available for a server, the applications generally prefer the IPv6 connection. However there are still lots of sites (like undeadly) which have better performance using v4. And in the case were v6 doesn't work, the way applications fall back to v4 is each time different and produce hard to understand messages. So often the easier solution is to disable IPv6 when you have the luck to have it :-(

    Comments
    1. By Anonymous Coward (82.95.158.67) on

      Hi,

      > One thing that bothers me: when both AAAA and A records are available for a server, the applications generally prefer the IPv6 connection. However there are still lots of sites (like undeadly) which have better performance using v4. And in the case were v6 doesn't work, the way applications fall back to v4 is each time different and produce hard to understand messages. So often the easier solution is to disable IPv6 when you have the luck to have it :-(
      >
      That's what a hosts(5) file is for. That and careful instructions in resolv.conf(5)

      Cheers.

      //

  11. By Steven Stillaway (samhain) steve@stillaway.net on

    Just checking if the IP is listed now. I think it was a firewalled DNS issue.

    Comments
    1. By Daniel Hartmeier (dhartmei) on

      > Just checking if the IP is listed now. I think it was a firewalled DNS issue.

      You mean the IP address shown by undeadly for a posted comment?

      For posts made by logged in users, comments show the user name instead of the IP(v6) address. If you want to see your IPv6 address, log out and post anonymously ;)

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]