I see fedoraproject.org has a AAAA record now (how new is that?) but the IP is not responding today. Access to fedoraproject was working up until today so I'm not sure if the AAAA record is brand new today or if it's just down.
$ ping6 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes ^C --- fedoraproject.org ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10721ms
I cannot use yum in this situation. I can access http://ipv6.google.com just fine.
On 08/28/2009 08:27 PM, Michael Cronenworth wrote:
I see fedoraproject.org has a AAAA record now (how new is that?) but the IP is not responding today. Access to fedoraproject was working up until today so I'm not sure if the AAAA record is brand new today or if it's just down.
$ ping6 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes ^C --- fedoraproject.org ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10721ms
I cannot use yum in this situation. I can access http://ipv6.google.com just fine.
try this.
Michael Cronenworth wrote:
I see fedoraproject.org has a AAAA record now (how new is that?) but the IP is not responding today. Access to fedoraproject was working up until today so I'm not sure if the AAAA record is brand new today or if it's just down.
The AAAA records were added in the past few hours. If it's causing any problems the folks in #fedora-admin on irc.freenode.net would probably like to know. That, or file a ticket with the infrastructure team via https://fedorahosted.org/fedora-infrastructure/ (this requires a Fedora Account).
$ ping6 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes ^C --- fedoraproject.org ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10721ms
I cannot use yum in this situation. I can access http://ipv6.google.com just fine.
Hmmm, I'm sure blocking access to updates was part of the IPv6 rollout plan. ;)
If the problem persists, filing a ticket or poking into #fedora-admin might be the quickest way to get it resolved. I'll point this thread out a while, in case it's a known problem.
Once upon a time, Michael Cronenworth mike@cchtml.com said:
I see fedoraproject.org has a AAAA record now (how new is that?) but the IP is not responding today. Access to fedoraproject was working up until today so I'm not sure if the AAAA record is brand new today or if it's just down.
$ ping6 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes ^C --- fedoraproject.org ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10721ms
It has been just over an hour since your message, but it works for me now:
$ ping6 -c1 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes 64 bytes from 2610:28:200:1::fed0:1: icmp_seq=1 ttl=56 time=57.6 ms
--- fedoraproject.org ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 57ms rtt min/avg/max/mdev = 57.627/57.627/57.627/0.000 ms
On Fri, 28 Aug 2009, Michael Cronenworth wrote:
I see fedoraproject.org has a AAAA record now (how new is that?) but the IP is not responding today. Access to fedoraproject was working up until today so I'm not sure if the AAAA record is brand new today or if it's just down.
$ ping6 fedoraproject.org PING fedoraproject.org(2610:28:200:1::fed0:1) 56 data bytes ^C --- fedoraproject.org ping statistics --- 11 packets transmitted, 0 received, 100% packet loss, time 10721ms
I cannot use yum in this situation. I can access http://ipv6.google.com just fine.
We had one of those deals where DNS was rolled out before the actual hosts came online with ipv6, please do let me know if you're still seeing this issue.
-Mike
Once upon a time, Todd Zullinger tmz@pobox.com said:
The AAAA records were added in the past few hours. If it's causing any problems the folks in #fedora-admin on irc.freenode.net would probably like to know. That, or file a ticket with the infrastructure team via https://fedorahosted.org/fedora-infrastructure/ (this requires a Fedora Account).
It appears that the mirrormanager is not handling IPv6 correctly. I have a Hurricane Electric IPv6 tunnel (from Ashburn, Virginia) to my home network, and now mirrormanager is giving me non-geolocated URLs. I needed to install something, and one hit downloaded from Brazil, and the next hit from Russia.
I've filed this here:
https://fedorahosted.org/fedora-infrastructure/ticket/1645
Chris Adams wrote:
It appears that the mirrormanager is not handling IPv6 correctly. I have a Hurricane Electric IPv6 tunnel (from Ashburn, Virginia) to my home network, and now mirrormanager is giving me non-geolocated URLs. I needed to install something, and one hit downloaded from Brazil, and the next hit from Russia.
I've filed this here:
I recall Matt Domsch mentioning this when he set things up yesterday. The geo-ip information available to mirrormanager from Maxmind isn't very complete.
Once upon a time, Todd Zullinger tmz@pobox.com said:
I recall Matt Domsch mentioning this when he set things up yesterday. The geo-ip information available to mirrormanager from Maxmind isn't very complete.
I was just looking to see if Maxmind had an IPv6 database (we use Maxmind's country database that is available for free here), and they show US for my IPv6 IP. I would guess there's a bug in mirrormanager's handling of IPv6 and geoip lookups.
Chris Adams wrote:
I was just looking to see if Maxmind had an IPv6 database (we use Maxmind's country database that is available for free here), and they show US for my IPv6 IP. I would guess there's a bug in mirrormanager's handling of IPv6 and geoip lookups.
Ahh, then yeah, that seems like a bug. I'll leave that to you and Matt in the ticket. :)
On Sat, 29 Aug 2009, Todd Zullinger wrote:
Chris Adams wrote:
I was just looking to see if Maxmind had an IPv6 database (we use Maxmind's country database that is available for free here), and they show US for my IPv6 IP. I would guess there's a bug in mirrormanager's handling of IPv6 and geoip lookups.
Ahh, then yeah, that seems like a bug. I'll leave that to you and Matt in the ticket. :)
For others with this issue, we've disabled ipv6 for mirrormanager for now until we get the seperate maxmind database up and ready.
-Mike
On 08/28/2009 08:44 PM, Mike McGrath wrote:
We had one of those deals where DNS was rolled out before the actual hosts came online with ipv6, please do let me know if you're still seeing this issue
I waited 24 hours and I still don't see a ping response or HTTP. It may be my end though since others are reporting it working.
$ traceroute6 fedoraproject.org traceroute to fedoraproject.org (2610:28:200:1::fed0:1), 30 hops max, 80 byte packets 1 2002:615d:dfb4:1::1 (2002:615d:dfb4:1::1) 1.965 ms 2.555 ms 2.443 ms 2 * * * 3 sl-bb1v6-sea-t-1001.sprintv6.net (2001:440:1239:1008::2) 75.719 ms 76.949 ms 76.893 ms 4 2001:440:1239:1013::2 (2001:440:1239:1013::2) 101.082 ms 100.954 ms 100.893 ms 5 * * * 6 * * * 7 * * * 8 * * * 9 * * * 10 * * * 11 * * * 12 * * * 13 * * * 14 * *^C
On Sat, 29 Aug 2009, Michael Cronenworth wrote:
On 08/28/2009 08:44 PM, Mike McGrath wrote:
We had one of those deals where DNS was rolled out before the actual hosts came online with ipv6, please do let me know if you're still seeing this issue
I waited 24 hours and I still don't see a ping response or HTTP. It may be my end though since others are reporting it working.
$ traceroute6 fedoraproject.org traceroute to fedoraproject.org (2610:28:200:1::fed0:1), 30 hops max, 80 byte packets 1 2002:615d:dfb4:1::1 (2002:615d:dfb4:1::1) 1.965 ms 2.555 ms 2.443 ms 2 * * * 3 sl-bb1v6-sea-t-1001.sprintv6.net (2001:440:1239:1008::2) 75.719 ms 76.949 ms 76.893 ms 4 2001:440:1239:1013::2 (2001:440:1239:1013::2) 101.082 ms 100.954 ms 100.893 ms 5 * * * 6 * * * 7 * * * 8 * * * 9 * * * 10 * * * 11 * * * 12 * * * 13 * * * 14 * *^C
When you get a moment can you stop by #fedora-admin on irc.freenode.net. So far this is the only issue reported but I'd like to follow up with it. The weekends are hit and miss but during the week I'm sure one of us involved in this can do better testing.
-Mike
Once upon a time, Michael Cronenworth mike@cchtml.com said:
I waited 24 hours and I still don't see a ping response or HTTP. It may be my end though since others are reporting it working.
$ traceroute6 fedoraproject.org traceroute to fedoraproject.org (2610:28:200:1::fed0:1), 30 hops max, 80 byte packets 1 2002:615d:dfb4:1::1 (2002:615d:dfb4:1::1) 1.965 ms 2.555 ms 2.443 ms 2 * * * 3 sl-bb1v6-sea-t-1001.sprintv6.net (2001:440:1239:1008::2) 75.719 ms 76.949 ms 76.893 ms 4 2001:440:1239:1013::2 (2001:440:1239:1013::2) 101.082 ms 100.954 ms 100.893 ms
I also have an IPv6 tunnel to Sprint, and I see it working:
traceroute6 to fedoraproject.org (2610:28:200:1::fed0:1) from 2001:440:eeee:ff4b::2, 64 hops max, 12 byte packets 1 sl-bb1v6-fw-t-37.sprintv6.net (2001:440:eeee:ff4b::1) 21.514 ms 21.615 ms 21.472 ms 2 sl-bb1v6-sea-t-1001.sprintv6.net (2001:440:1239:1008::2) 88.558 ms 88.986 ms 88.611 ms 3 2001:440:1239:1013::2 (2001:440:1239:1013::2) 112.097 ms 112.111 ms 128.273 ms 4 ge-0-1-0.82.rtr.salt.net.internet2.edu (2001:468:ff:17c4::1) 139.529 ms 140.165 ms 139.725 ms 5 * * * 6 * * * 7 * * * 8 2001:468:ffff:155::2 (2001:468:ffff:155::2) 143.351 ms 161.345 ms 161.480 ms 9 2610:28:10e:2::1 (2610:28:10e:2::1) 146.344 ms 145.950 ms 147.411 ms 10 2610:28:105:13::2 (2610:28:105:13::2) 145.395 ms 148.297 ms 146.027 ms 11 2610:28:200:1::fed0:1 (2610:28:200:1::fed0:1) 146.802 ms !P 150.567 ms !P 150.150 ms !P
On Sat, Aug 29, 2009 at 08:51:56AM -0500, Chris Adams wrote:
Once upon a time, Todd Zullinger tmz@pobox.com said:
The AAAA records were added in the past few hours. If it's causing any problems the folks in #fedora-admin on irc.freenode.net would probably like to know. That, or file a ticket with the infrastructure team via https://fedorahosted.org/fedora-infrastructure/ (this requires a Fedora Account).
It appears that the mirrormanager is not handling IPv6 correctly. I have a Hurricane Electric IPv6 tunnel (from Ashburn, Virginia) to my home network, and now mirrormanager is giving me non-geolocated URLs. I needed to install something, and one hit downloaded from Brazil, and the next hit from Russia.
I've filed this here:
I've rolled out a new version of MirrorManager that uses very new MaxMind GeoIP code to do IPv6 lookups, and 6to4 (IPv4) lookups too. This is in use now. Please let me know if you experience any further problems.
The MaxMind GeoIP IPv6 database only has a couple thousand entries, so it's bound to have gaps. The 6to4 lookups will help close that to some extent. But where a user isn't using 6to4, and their network isn't listed in the database already, you'll get the global list back.
Thanks, Matt
On 09/11/2009 12:29 AM, Matt Domsch wrote:
I've rolled out a new version of MirrorManager that uses very new MaxMind GeoIP code to do IPv6 lookups, and 6to4 (IPv4) lookups too. This is in use now. Please let me know if you experience any further problems.
The MaxMind GeoIP IPv6 database only has a couple thousand entries, so it's bound to have gaps. The 6to4 lookups will help close that to some extent. But where a user isn't using 6to4, and their network isn't listed in the database already, you'll get the global list back.
Great news. My 6to4 connection is still working (*crosses fingers*) so I hope to be able to test this.
Now you need to get redhat.com on IPv6. ;)
Once upon a time, Matt Domsch Matt_Domsch@dell.com said:
The MaxMind GeoIP IPv6 database only has a couple thousand entries, so it's bound to have gaps.
If it has more than 1870, then it probably is good. That's how many active IPv6 routes there are in my (default-free) BGP IPv6 routing table right now. Most ISPs (even many of the largest) only have one prefix in use, at least right now. Also, the IPv6 space is supposed to be less fragmented when it comes to location.
-
Chris Adams -
Jim -
Matt Domsch -
Michael Cronenworth -
Mike McGrath -
Todd Zullinger