IPv6 beta for Bigpipe

Forums › Spark New Zealand (including Skinny and BigPipe) › IPv6 beta for Bigpipe

1 | ... | 23 | 24 | 25 | 26 | 27 | 28 | 29

51 posts

Master Geek

#1948372 29-Jan-2018 22:37

For the record, I now have a working, dynamic configuration. The DHCP server is still very moody and only rarely responds to solicitations, but when it does, the following make things work. ppp0 is my PPPoE interface, and eth0 is on the LAN.

 
--[ /etc/wide-dhcpv6/dhcp6c.conf ]-------------------------------------------- 
profile default 
{ 
  request domain-name-servers; 
 
  script "/etc/wide-dhcpv6/dhcp6c-script"; 
}; 
 
interface ppp0 {  
    send ia-pd 0; 
    #send rapid-commit; 
}; 
  
id-assoc pd 0 { 
  prefix-interface eth0 { 
    sla-id 0; 
    sla-len 4; 
    ifid 1; 
  }; 
}; 
 
--[ /etc/radvd.conf ]--------------------------------------------------------- 
interface eth0 
{ 
   AdvSendAdvert on; 
   AdvDefaultPreference high; 
   prefix ::/64 
   { 
   }; 
}; 
 
--[ cat /etc/ppp/ip-{up,down}.d/local-dhcp6c ]-------------------------------- 
#!/bin/sh 
set -eu 
 
case "$0" in 
  (*-down*) MODE=stop;; 
  (*) MODE=start;; 
esac 
 
env IFACE=$PPP_IFACE MODE=$MODE /etc/wide-dhcpv6/dhcp6c-ifupdown 
 
case "$MODE" in 
  (start) 
    ip -6 route add default dev $PPP_IFACE 
    systemctl restart radvd.service 
    ;; 
  (stop) 
    systemctl stop radvd.service 
    ;; 
esac

madduck

51 posts

Master Geek

#1949882 1-Feb-2018 09:44

So this worked for a whole day, then the DSL connection dropped and since then (2 days now), the DHCPv6 client has been sending out solicitations without getting any response.

I've been given a prefix previously. However, if I now statically assign this prefix (please $DEITY forgive…) as dhcp6c had in the past, then I still don't get v6 connectivity, even though the link seems to be established:

 
root@atom:/# ping6 ff02::2%ppp0 
PING ff02::2%ppp0(ff02::2) 56 data bytes 
64 bytes from fe80::e2ac:f1ff:fe81:b104: icmp_seq=1 ttl=64 time=16.3 ms 
[…] 
^C 
 
root@atom:/# ip -6 a 
1: lo:  mtu 65536 state UNKNOWN qlen 1 
    inet6 ::1/128 scope host  
       valid_lft forever preferred_lft forever 
2: eth0:  mtu 1500 state UP qlen 1000 
    inet6 2403:9800:c048:37d0::1/64 scope global  
       valid_lft forever preferred_lft forever 
    inet6 fe80::ba27:ebff:fe5d:5c08/64 scope link  
       valid_lft forever preferred_lft forever 
3: eth1:  mtu 1500 state UP qlen 1000 
    inet6 fe80::250:b6ff:fe17:8546/64 scope link  
       valid_lft forever preferred_lft forever 
44: ppp0:  mtu 1492 state UNKNOWN qlen 3 
    inet6 fe80::4:2:89/10 scope link  
       valid_lft forever preferred_lft forever 
 
root@atom:/# ip -6 r 
2403:9800:c048:37d0::/64 dev eth0  proto kernel  metric 256  
fe80::/64 dev eth1  proto kernel  metric 256  
fe80::/64 dev eth0  proto kernel  metric 256  
fe80::/10 dev ppp0  metric 1  
fe80::/10 dev ppp0  proto kernel  metric 256  
default dev ppp0  metric 1024  
 
root@atom:/# ping6 -nc3 debian.org 
PING debian.org(2001:41c8:1000:21::21:4) 56 data bytes 
 
--- debian.org ping statistics --- 
3 packets transmitted, 0 received, 100% packet loss, time 2005ms

Any ideas? Ideally, the DHCP server gets fixed, but do you know of workarounds until that has happened?

jfp

7 posts

Wannabe Geek

#1949891 1-Feb-2018 09:57

By assigning it statically do you mean changing radvd.conf?

Ie:

interface enp1s0
{
AdvSendAdvert on;
MinRtrAdvInterval 30;
MaxRtrAdvInterval 100;
AdvDefaultPreference low;
#A subnet from the /60 allocation
prefix BIGPIPESUBNET/64
{
AdvOnLink on;
AdvAutonomous on;
AdvRouterAddr on;
};
};

madduck

51 posts

Master Geek

#1949914 1-Feb-2018 10:33

No, radvd is a whole other beast and takes care of advertising any prefixes to the LAN (SLAAC).

Right now I am trying to get the router to be able to speak IPv6 itself, for a start.

There's a default route in place, and the router can ping6 the remote link-local IPv6 address pppd received:

 
root@atom:/# ping6 -nc1 -I ppp0 fe80::e2ac:f1ff:fe81:b104 
PING fe80::e2ac:f1ff:fe81:b104(fe80::e2ac:f1ff:fe81:b104) from fe80::4:2:89 ppp0: 56 data bytes 
64 bytes from fe80::e2ac:f1ff:fe81:b104: icmp_seq=1 ttl=64 time=13.5 ms 
 
--- fe80::e2ac:f1ff:fe81:b104 ping statistics --- 
1 packets transmitted, 1 received, 0% packet loss, time 0ms 
rtt min/avg/max/mdev = 13.581/13.581/13.581/0.000 ms

So now I add a valid, global IPv6 address with the prefix that I had been assigned the last time the DHCP server replied, to either ppp0 or eth0 (either would work):

ip a a 2403:9800:c048:37d0::1/64 dev ppp0

But I cannot use that address, i.e. ping6 to the outside doesn't work (cf. the failing ping6 to debian.org quoted above).

jfp

7 posts

Wannabe Geek

#1949948 1-Feb-2018 10:53

The interesting thing is that my setup works, with that radvd.conf, but the only public ipv6 addresses are on my LAN, assigned by the radvd.

My WAN and ppp0 only have fe80:: addresses, but the router can connect to the ipv6 internet using it's LAN ipv6 address.

Try manually adding a public /64 ipv6 address to you LAN and it will probably work.

madduck

51 posts

Master Geek

#1950022 1-Feb-2018 12:21

radvd correctly advertises the prefix to the LAN, and while they can all ping the router merily, they too cannot access the outer 'Net using IPv6.

So given the situation that link-local v6 works between my PPPoE endpoint and the Bigpipe router, that their DHCP server seems wonky, and that the previously assigned prefix cannot be used as a globally-scoped address, or at least isn't being forwarded by them, I am fairly sure that I won't be able to fix this by poking at bits over here alone.

Does anyone at Bigpipe actually still follow this thread?

madduck

51 posts

Master Geek

#1950251 1-Feb-2018 17:44

Ping @JoshBigpipe @ipv6pipe ?

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

madduck

51 posts

Master Geek

#1952121 5-Feb-2018 17:16

Something changed this morning, and now the DHCP server replies reliably to pretty much every request. Thank you @ipv6pipe for doing whatever you've been doing.

matisyahu

1623 posts

Uber Geek

Trusted

#1953992 9-Feb-2018 02:52

madduck: Something changed this morning, and now the DHCP server replies reliably to pretty much every request. Thank you @ipv6pipe for doing whatever you've been doing.

I've enabled ipv6 on my Google Wifi, I tested my connection and it appears that BigPipe has switched on ipv6 for everyone (I never opted in for the beta testing of ipv6).

"When the people are being beaten with a stick, they are not much happier if it is called 'the People's Stick'"

Aaroona

3196 posts

Uber Geek

#1986145 30-Mar-2018 17:47

I assume IPv6 is still not production released yet? Is there an ETA on it?

My IPv6 has stopped working (though I believe it's possibly related to some IP changes the support team did to my connection). I've PM's the ipv6pipe account, so hopefully I'll get something from there...

Aaroona

3196 posts

Uber Geek

#1986867 1-Apr-2018 17:49

Aaroona:

I assume IPv6 is still not production released yet? Is there an ETA on it?

My IPv6 has stopped working (though I believe it's possibly related to some IP changes the support team did to my connection). I've PM's the ipv6pipe account, so hopefully I'll get something from there...

Some additional info.... It seems I am getting my /60 delegation and am getting an IPv6 address to my devices, but for whatever reason, it doesn't appear a default gateway is coming through. This is the same issue I was having with IPv4 the other day and they had to change my IP address apparently to make it work.

trying wget commands from my router, forcing IPv6 either comes back with destination host unreachable or no route found type errors.

michaelmurfy

meow
13243 posts

Uber Geek

Moderator

ID Verified

Trusted

Lifetime subscriber

#1986868 1-Apr-2018 17:53

@Aaroona Working with some routers I've got configured:

:~$ ping6 google.com
PING google.com(syd09s13-in-x0e.1e100.net) 56 data bytes
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=1 ttl=54 time=35.2 ms
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=2 ttl=54 time=34.4 ms
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=3 ttl=54 time=34.5 ms
^C
--- google.com ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 34.485/34.746/35.254/0.389 ms

This is a misconfiguration on your end - what router are you using?

Michael Murphy | https://murfy.nz
Referral Links: Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.

Aaroona

3196 posts

Uber Geek

#1986872 1-Apr-2018 17:57

michaelmurfy:

@Aaroona Working with some routers I've got configured:

:~$ ping6 google.com
PING google.com(syd09s13-in-x0e.1e100.net) 56 data bytes
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=1 ttl=54 time=35.2 ms
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=2 ttl=54 time=34.4 ms
64 bytes from syd09s13-in-x0e.1e100.net: icmp_seq=3 ttl=54 time=34.5 ms
^C
--- google.com ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 34.485/34.746/35.254/0.389 ms

This is a misconfiguration on your end - what router are you using?

The same router I've been using with IPv6 for the last year - WNDR3700 running LEDE. Only started having issues the other day when Bigpipe made changes to my account, which included messing with my Static IP and putting me behind CGNAT, then me requesting a move back, more issues, then finally resolved when they issued me a new, different static IP address. since then, IPv6 has been borked.

All I get when I try ping from my router:

root@LEDE:~# ping6 google.co.nz
PING google.co.nz (2404:6800:4006:804::2003): 56 data bytes

Doesn't do a thing.
Wget from the router and RPi comes up like this at the moment;

root@LEDE:~# wget --inet6-only https://geekzone.co.nz
--2018-04-01 17:56:30-- https://geekzone.co.nz/
Resolving geekzone.co.nz... 2400:cb00:2048:1::6818:20e, 2400:cb00:2048:1::6818:30e
Connecting to geekzone.co.nz|2400:cb00:2048:1::6818:20e|:443... failed: Host is unreachable.
Connecting to geekzone.co.nz|2400:cb00:2048:1::6818:30e|:443... failed: Host is unreachable.

EDIT:

Some additional info:

Type: dhcpv6-pd
Prefix Delegated: 2403:XXXX:c048:300::/60
Address: 2403:XXXX:c041:ffe0:XXX:3246:XXXX:5908/128
Gateway: fe80::e2ac:f1ff:fe88:a202
DNS 1: 2403:9800:c041:1000::a
DNS 2: 2403:9800:c041:1000::b
Connected: 0h 51m 21s

I've blanked out a couple of bits to mask, but should be enough to see my point, re the gateway.

Aaroona

3196 posts

Uber Geek

#1988314 4-Apr-2018 17:19

Just to update. There was a provisioning issue on their side. Looks like they turned off my IPv6 when they updated my IPv4 settings for whatever reason.
Spoke to support via online chat and it's all up and running again. A matter of seconds.

All working again and surfing on v6! :D

ccar063

12 posts

Geek

#1997818 17-Apr-2018 13:39

matisyahu:

I've enabled ipv6 on my Google Wifi, I tested my connection and it appears that BigPipe has switched on ipv6 for everyone (I never opted in for the beta testing of ipv6).

It does appear that they're rolling it out to a wider audience. ipv6 is also enabled on my account and I haven't opted in for the beta testing.

1 | ... | 23 | 24 | 25 | 26 | 27 | 28 | 29