User comments on ISPs
  >> Zen Internet


Register (or login) on our website and you will not see this ad.


  Print Thread
Standard User planetf1
(experienced) Wed 23-Dec-20 08:27:38
Print Post

Dual stack IPv4/6


[link to this post]
 
(Kept getting blocked by cloudflare ray id Cloudflare Ray ID: 6060b2f23a250732. Seems a mention of e-t-c h-o-s-t-s in the text caused it to fail? ).


Hi,
I have just switched to Zen with full dual stack IPv4/6 support, using a Fritzbox 7530 and am experiencing a delay , seemingly in name resolution, when using 'ping6'.

The same configuration is also supporting iOS, windows, linux - and these devices all seem to work fine.

The symptom is:
'ping www.ibm.com' -> instant response, start getting icmp request/response
ping6 'www.ibm.com' -> About a 3-4s delay after typing the command before the pings start

ie:
$ ping6 www.ibm.com
... delay here ...
PING6(56=40+8+8 bytes) 2a02:8010:687f:0:2dbc:1ea2:a301:b111 --> 2a02:26f0:e8:491::b3a
16 bytes from 2a02:26f0:e8:491::b3a, icmp_seq=0 hlim=58 time=11.621 ms
16 bytes from 2a02:26f0:e8:491::b3a, icmp_seq=1 hlim=58 time=10.486 ms
16 bytes from 2a02:26f0:e8:491::b3a, icmp_seq=2 hlim=58 time=11.994 ms

The same behaviour is seen for other IPv6 sites such as ipv6.google.com

Using 'nslookup' I also get a delay:

$ nslookup
... delay actually occurs here
> ipv6.google.com
Server: 192.168.178.1
Address: 192.168.178.1#53

Non-authoritative answer:
ipv6.google.com canonical name = ipv6.l.google.com.
>

That delay is almost as if it's reverse resolving itself - or a localhost issue?

All the ipv6 testers I've tried in a browser suggest my IPv6 config is fine, sites reachable
I don't notice delays in google chrome
If IPv4 is disabled no delays occur
The router is set to automatic IPv6 (with fast commit)
The macOS Wifi adapter is set to automatic for ipv6 (and dhcp for ipv4)

scutil --dns ends with

DNS configuration (for scoped queries)

resolver #1
search domain[0] : fritz.box
nameserver[0] : 192.168.178.1
nameserver[1] : fd00::2e91:abff:fe55:2d26
if_index : 5 (en0)
flags : Scoped, Request A records, Request AAAA records
reach : 0x00020002 (Reachable,Directly Reachable Address)

So ipv6 addresses are being requested


##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting. Do not change this entry.
##
127.0.0.1 localhost
::1 localhost
fe80::1%lo0 localhost
# End of section
192.168.178.48 pi4

Any ideas why the delay?

Once past that ipv6 and ipv4 are both working fine it seems (from limited tests so far)

Edited by planetf1 (Wed 23-Dec-20 08:32:59)

Standard User planetf1
(experienced) Wed 13-Jan-21 14:30:11
Print Post

Re: Dual stack IPv4/6


[re: planetf1] [link to this post]
 
I should add I didn't see any issues outside of this - ie other systems all fine, as indeed was the mac other than ping6

For whatever reason, but almost definately a config change/update on my mac, even ping6 is working just fine now.
Standard User ukwiz
(fountain of knowledge) Wed 13-Jan-21 20:04:42
Print Post

Re: Dual stack IPv4/6


[re: planetf1] [link to this post]
 
I am on Zen with dual stack, and get a delay also get a delay - 9 seconds.

David

BT (poor) -> Zen (excellent) -> O2 (started well, went downhill -> IDNet (No complaints - but 100GB cap) -> Zen (gone downhill)


Register (or login) on our website and you will not see this ad.

Standard User Andrue
(eat-sleep-adslguide) Thu 14-Jan-21 21:50:15
Print Post

Re: Dual stack IPv4/6


[re: planetf1] [link to this post]
 
I'm with IDNet and last night whilst I got an immediate DNS look up I wasn't getting any replies. Curiously I wasn't getting any replies from any other external IPv6 addresses either although I was able to browse to IPv6 sites.

Tonight I'm getting replies and all seems normal.

C:\Users\Andrue>ping www.ibm.com

Pinging e2874.dscx.akamaiedge.net [2a02:26f0:e8:486::b3a] with 32 bytes of data:
Reply from 2a02:26f0:e8:486::b3a: time=12ms
Reply from 2a02:26f0:e8:486::b3a: time=12ms
Reply from 2a02:26f0:e8:486::b3a: time=13ms

Since we're using different ISPs it probably means nothing but I thought I'd mention it.

---
Andrue Cope
Brackley, UK

Edited by Andrue (Thu 14-Jan-21 21:51:07)

Standard User caffn8me
(eat-sleep-adslguide) Fri 15-Jan-21 00:27:22
Print Post

Re: Dual stack IPv4/6


[re: ukwiz] [link to this post]
 
In reply to a post by planetf1:
I have just switched to Zen with full dual stack IPv4/6 support, using a Fritzbox 7530 and am experiencing a delay , seemingly in name resolution, when using 'ping6'.
I'm also getting strange results from a Zen FTTC dual stack connection. I'd assumed it was related to a recent firewall firmware upgrade but it may not be and I'm away from the site at the moment. I won't downgrade the firewall remotely in case it falls over.

I've only recently enabled IPv6 and haven't done extensive testing - but it was working better a few weeks ago than it is now.
In reply to a post by ukwiz:
I am on Zen with dual stack, and get a delay also get a delay - 9 seconds.
If I do an IPv6 ping from the router directly - on the /64 address block I get 100% replies after a delay for the first dns lookup
caffeine-gw#ping ipv6.google.com
Translating "ipv6.google.com"...domain server (82.68.85.106)

Translating "ipv6.google.com"...domain server (82.68.85.106)

Translating "ipv6.google.com"...domain server (82.68.85.106) [OK]

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 2A00:1450:4009:816::200E, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 4/5/8 ms
The DNS lookup delay is about eight or nine seconds, so consistent with what you see.

When I do the ping immediately again, the DNS server (which is local) returns the cached result for ipv6.google.com and there's no delay;
caffeine-gw#ping ipv6.google.com
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 2A00:1450:4009:816::200E, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 4/5/8 ms
If I ping from the dual-stack DNS server on the /48 subnet behind the firewall there's no reply to the first seven ping requests. Then no more replies until 13 seconds, after which it only responds to 50% of pings.
espresso:~$ ping ipv6.google.com
PING ipv6.google.com(lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e)) 56 data bytes
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=8 ttl=114 time=5.71 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=13 ttl=114 time=5.75 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=15 ttl=114 time=5.49 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=17 ttl=114 time=5.51 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=19 ttl=114 time=6.00 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=21 ttl=114 time=26.4 ms
64 bytes from lhr25s25-in-x0e.1e100.net (2a00:1450:4009:816::200e): icmp_seq=23 ttl=114 time=5.78 ms
^C
--- ipv6.google.com ping statistics ---
26 packets transmitted, 7 received, 73.0769% packet loss, time 30000ms
rtt min/avg/max/mdev = 5.494/8.670/26.449/7.259 ms
The seven missing responses before the first reply are not due to a DNS lookup delay as this is the DNS server which was queried by the router - and is returning cached results. The DNS lookup itself is instant once the answer has been cached.

This suggests that the trouble is with the /48 routed subnet, possibly something at Zen's end.

If I traceroute from the router to ipv6.google.com I get this;
caffeine-gw#traceroute 2A00:1450:4009:816::200E
Type escape sequence to abort.
Tracing the route to lhr25s25-in-x0e.1e100.net (2A00:1450:4009:816::200E)

  1 2A02:8010::4:0:0:6 4 msec 4 msec 4 msec
  2 2A02:8010:0:404::16 8 msec 4 msec 4 msec
  3 2A02:8010:0:700::12 4 msec 4 msec 8 msec
I can determine that the first hop on Zen's network is 2A02:8010::4:0:0:6.

I can ping this happily from the router;
caffeine-gw#ping 2A02:8010::4:0:0:6
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 2A02:8010::4:0:0:6, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 4/4/4 ms
No delay, 100% reply.

I can ping my router's external IPv6 interface from my nameserver with no trouble and get no delay or packet loss.
espresso:~$ ping 2a02:8011:d000:::xxx:
PING 2a02:8011:d000:::xxx:(2a02:8011:d000:::xxx:) 56 data bytes
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=1 ttl=63 time=1.06 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=2 ttl=63 time=1.04 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=3 ttl=63 time=1.03 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=4 ttl=63 time=1.02 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=5 ttl=63 time=0.987 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=6 ttl=63 time=1.40 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=7 ttl=63 time=1.35 ms
64 bytes from 2a02:8011:d000:::xxx: icmp_seq=8 ttl=63 time=1.03 ms
^C
--- 2a02:8011:d000:::xxx: ping statistics ---
8 packets transmitted, 8 received, 0% packet loss, time 7012ms
rtt min/avg/max/mdev = 0.987/1.113/1.397/0.151 ms
When I try to ping IPv6 beyond my own router to the first hop I get this;
espresso:~$ ping 2A02:8010::4:0:0:6
PING 2A02:8010::4:0:0:6(2a02:8010::4:0:0:6) 56 data bytes
64 bytes from 2a02:8010::4:0:0:6: icmp_seq=44 ttl=253 time=5.96 ms
64 bytes from 2a02:8010::4:0:0:6: icmp_seq=46 ttl=253 time=5.65 ms
64 bytes from 2a02:8010::4:0:0:6: icmp_seq=48 ttl=253 time=5.64 ms
64 bytes from 2a02:8010::4:0:0:6: icmp_seq=50 ttl=253 time=23.8 ms
^C
--- 2A02:8010::4:0:0:6 ping statistics ---
66 packets transmitted, 4 received, 93.9394% packet loss, time 67579ms
rtt min/avg/max/mdev = 5.641/10.273/23.845/7.836 ms
Note that the first reply in this case is sequence 44. This number changes each time (I've seen >70) , and the packet loss after the first reply is also variable but at least every other packet. All the time, the router can ping with no packet loss.

Those are my experiences so far.

Sarah

--
If I can't drink my bowl of coffee three times daily, then in my torment, I will shrivel up like a piece of roast goat

Spiders on coffee - Badass spiders on drugs
Standard User prlzx
(experienced) Fri 15-Jan-21 13:44:54
Print Post

Re: Dual stack IPv4/6


[re: planetf1] [link to this post]
 
No delay in external DNS resolution here on dual-stack Zen.

I do have another router in my LAN acting as primary DNS server (*) but it forwards non-local queries to the fritz.box which remains as the default gateway for v4/v6.
However if I direct queries from PC to the fritz.box instead responses continue to be instant.

* because the other router has VPN duties and knows about LAN and Intranet DNS namespaces that fritz would not



prlzx on Zen: FTTC (VDSL) at ~40Mbps / 10Mbps
with IP4/6 (no v6? - not true Internet)
Standard User caffn8me
(eat-sleep-adslguide) Fri 15-Jan-21 16:22:55
Print Post

Re: Dual stack IPv4/6


[re: prlzx] [link to this post]
 
In reply to a post by prlzx:
No delay in external DNS resolution here on dual-stack Zen.

I do have another router in my LAN acting as primary DNS server (*) but it forwards non-local queries to the fritz.box which remains as the default gateway for v4/v6.
However if I direct queries from PC to the fritz.box instead responses continue to be instant.

* because the other router has VPN duties and knows about LAN and Intranet DNS namespaces that fritz would not
Just a question; is this from the /64 or the /48?

Sarah

--
If I can't drink my bowl of coffee three times daily, then in my torment, I will shrivel up like a piece of roast goat

Spiders on coffee - Badass spiders on drugs
Standard User prlzx
(experienced) Sun 17-Jan-21 05:27:39
Print Post

Re: Dual stack IPv4/6


[re: caffn8me] [link to this post]
 
In reply to a post by caffn8me:
Just a question; is this from the /64 or the /48?


When I override the DNS lookups from the PC to use the fritz.box like anything else I access on the LAN I am generally using the hostname or fqdn rather than IP addresses, so

dig www.example.net @fritz.box

For IPv6 the LAN address (global scope) of the fritz.box is automatically the first (0th) /64 subnet out of the delegated /48 prefix.
For the host part of the address it is using the address derived from the MAC of its LAN interface.

However the address of fritz.box as you might expect itself resolves to
- its LAN address using IPv6 ( global scope )
- its LAN address using IPv6 ( ULA starting fd00: )
- its LAN address using IPv4

the same dig works when the target of the @ is IPv4 address, the IPv6 global scope, the IPv6 ULA scope
and (if properly qualified with the interface) the IPv6 link-local address.

The unrelated external /64 on the Internet facing interface is not used for LAN DNS lookups.



prlzx on Zen: FTTC (VDSL) at ~40Mbps / 10Mbps
with IP4/6 (no v6? - not true Internet)

Edited by prlzx (Sun 17-Jan-21 05:57:19)

  Print Thread

Jump to