Dotaz: problem so spolahlivostou IPv6 na routeroch

[disposable]

Mám niekoľko Ubuntu 10.04 boxov, na ktorých beží Quagga a používam ich ako BGP routery. V IPv4 beží všetko ako má, latencia je nízka a packet loss je 0 percent pri všetkých peer-och. napr.:

root@r6-REDBUS:/etc/network# ping 213.1XX.79.197
PING 213.1XX.79.197 (213.1XX.79.197) 56(84) bytes of data.
64 bytes from 213.1XX.79.197: icmp_seq=1 ttl=64 time=0.939 ms
64 bytes from 213.1XX.79.197: icmp_seq=2 ttl=64 time=1.22 ms
64 bytes from 213.1XX.79.197: icmp_seq=3 ttl=64 time=0.927 ms
64 bytes from 213.1XX.79.197: icmp_seq=4 ttl=64 time=1.19 ms
64 bytes from 213.1XX.79.197: icmp_seq=5 ttl=64 time=2.34 ms
64 bytes from 213.1XX.79.197: icmp_seq=6 ttl=64 time=0.968 ms
64 bytes from 213.1XX.79.197: icmp_seq=7 ttl=64 time=1.15 ms
64 bytes from 213.1XX.79.197: icmp_seq=8 ttl=64 time=0.869 ms
64 bytes from 213.1XX.79.197: icmp_seq=9 ttl=64 time=0.955 ms
64 bytes from 213.1XX.79.197: icmp_seq=10 ttl=64 time=0.896 ms

 

Ten istý peer v IPv6, ale ukáže toto: (väčšinou, ale ani nedokážem pingnúť, dostanem iba "Network is unreachable")

root@r6-REDBUS:br />PING 2001:438:fXXX::4b9(2001:438:fXXX::4b9) from 2001:438:fXXX::4ba bond1.849: 56 data bytes
From 2001:438:fXXX::4ba icmp_seq=1 Destination unreachable: Address unreachable
From 2001:438:fXXX::4ba icmp_seq=2 Destination unreachable: Address unreachable
From 2001:438:fXXX::4ba icmp_seq=3 Destination unreachable: Address unreachable
ping: sendmsg: Network is unreachable
ping: sendmsg: Network is unreachable
From 2001:438:fXXX::4ba icmp_seq=4 Destination unreachable: Address unreachable
ping: sendmsg: Network is unreachable
64 bytes from 2001:438:fXXX::4b9: icmp_seq=8 ttl=64 time=653 ms
64 bytes from 2001:438:fXXX::4b9: icmp_seq=9 ttl=64 time=7.24 ms
64 bytes from 2001:438:fXXX::4b9: icmp_seq=10 ttl=64 time=33.9 ms
64 bytes from 2001:438:fXXX::4b9: icmp_seq=11 ttl=64 time=1.19 ms

^C
--- 2001:438:fXXX::4b9 ping statistics ---
61 packets transmitted, 4 received, +4 errors, 93% packet loss, time 60150ms
rtt min/avg/max/mdev = 1.190/173.867/653.093/276.955 ms

 

 

IPv4 aj IPv6 k tomuto Peerovi ide cez ten istý kábel, po tej istej vlan. Sieť je v /etc/network/interfaces nakonfigurovaná nasledovne:

auto bond1.849
iface bond1.849 inet static
address 213.XXX.79.198
netmask 255.255.255.252
network 213.XXX.79.196
broadcast 213.XXX.79.199
vlan_raw_interface bond1
up /sbin/ip -6 addr add 2001:438:fXXX::4ba/64 dev bond1.849
down /sbin/ip -6 addr del 2001:438:fXXX::4ba/64 dev bond1.849

root@r6-REDBUS:/etc/network# ip -6 link show dev bond1.849
23: bond1.849@bond1: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP
link/ether 00:10:f3:1a:48:5a brd ff:ff:ff:ff:ff:ff
root@r6-REDBUS:/etc/network# ip -6 addr show dev bond1.849
23: bond1.849@bond1: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500
inet6 2001:438:fXXX::4ba/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::210:f3ff:fe1a:485a/64 scope link
valid_lft forever preferred_lft forever



 

 

Tento problém mám so všetkými peermi na všetkých routroch. Vo všetkých prípadoch IPv4 funguje bezchybne, IPv6 iba sporadicky. Väčšinou mám striedavo "Network is unreachable" a "Address unreachable". Po niekoľkých pokusoch začne ping6 na chvíľu fungovať, ale dosť nepravidelne

root@r6-REDBUS:/etc/network# ping6 2001:1900:XXX:2:2::631
connect: Network is unreachable
root@r6-REDBUS:/etc/network# ping6 2001:1900:XXX:2:2::631
connect: Network is unreachable
root@r6-REDBUS:/etc/network# ping6 2001:1900:XXX:2:2::631
connect: Network is unreachable
root@r6-REDBUS:/etc/network# ping6 2001:1900:XXX:2:2::631
connect: Network is unreachable
root@r6-REDBUS:/etc/network# ping6 2001:1900:XXX:2:2::631
PING 2001:1900:XXX:2:2::631(2001:1900:XXX:2:2::631) 56 data bytes
From 2001:1900:XXX:2:2::632 icmp_seq=1 Destination unreachable: Address unreachable
From 2001:1900:XXX:2:2::632 icmp_seq=2 Destination unreachable: Address unreachable
From 2001:1900:XXX:2:2::632 icmp_seq=3 Destination unreachable: Address unreachable
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=4 ttl=64 time=2218 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=5 ttl=64 time=1218 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=6 ttl=64 time=218 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=7 ttl=64 time=135 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=8 ttl=64 time=39.1 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=9 ttl=64 time=7.51 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=10 ttl=64 time=19.0 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=11 ttl=64 time=1.06 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=12 ttl=64 time=0.826 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=13 ttl=64 time=14.4 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=14 ttl=64 time=28.3 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=15 ttl=64 time=1.50 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=16 ttl=64 time=16.5 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=17 ttl=64 time=339 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=18 ttl=64 time=242 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=19 ttl=64 time=406 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=20 ttl=64 time=196 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=21 ttl=64 time=10.3 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=22 ttl=64 time=182 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=23 ttl=64 time=49.2 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=24 ttl=64 time=11.3 ms
64 bytes from 2001:1900:XXX:2:2::631: icmp_seq=25 ttl=64 time=1.82 ms



 

link k peer-ovi je router ---> juniper ex4200 switch ---> virtual cable provider (napr.: gt-t.net) ---> peer router. Link od môjho routera až po peerov router ide po dedikovanej VLAN. Juniper switch na mojej strane je nakonfigurovaný takto (port idúci k virtual cable provider):


ge-2/0/19 {
ether-options {
no-auto-negotiation;
link-mode full-duplex;
speed {
1g;
}
}
unit 0 {
description "Packet Exchange Peering Cable";
family ethernet-switching {
port-mode trunk;
vlan {
members [ TATA-BGP-3040 PACKETEXCHANGE-EXCHANGEPOINT-502 ABOVENET-BGP-849 ];
}
native-vlan-id 80;
}
}
}

 

port idúci k môjmu routeru:

ge-2/0/5 {
description R6-eth2;
unit 0 {
family ethernet-switching {
port-mode trunk;
vlan {
members [ RED-INTERNAL-96 RED-INTERNAL-69 PACKETEXCHANGE-EXCHANGEPOINT-502 TATA-BGP-3040 LEVEL3-BGP-80 ABOVENET-BGP-849 DATAHOP-TRANSIT-640 LINX-EXTREME-528 ];
}
}
}
}



Chápem že tento opis je dosť zdĺhavý, ale celkovo je to celkom jednoduchý setup. Bol by som vďačný, keby mi niekto vysvetlil kde robím chybu. Nemám poňatia či mám zle nastavené sieťové rozhranie na routeri, či treba nastaviť špecifické MTU na routeri (alebo na switchi), alebo kde som to vlastne posral.

čo sa týka ipv6 sysctl nastavení, tak mám iba:
net.ipv6.conf.all.forwarding=1
#net.ipv6.conf.all.accept_redirects = 0
net.ipv6.conf.all.accept_source_route = 1

Odpovědi

[pavlix]

Na to je potřeba sledovat, co máš skutečně ve směrovací tabulce. Jinde bych asi problém nehledal. Popis je opravdu příliš zdlouhavý a není v něm izolováno, co konkrétně nefunguje či dělá problémy.

[disposable]

no kedze je to BGP router s roznymi transit peermi, tak v tabulke je cely internet (a niekolko krat). Pre uvedeny problem je to ale nepodstatne. Ja len potrebujem vediet preco mi ping6 nefunguje a ked funguje, tak preco len tak nespolahlivo. Toto je moj prvy pokus o pracu s IPv6 a mozno mi uniklo nieco uplne primitivne.

[pavlix]

Je mi líto, ale bez podstatných informací nemůžu pomoci.

[NN]

Jsou cestou jeste nejake aktivni prvky ?

[disposable]

v našom datacentre je to len

náš router ----> náš switch ----> switch poskytovateľa konektivity, cez ktorého peerujem s transit providermi (komunikácia s každým peerom je cez osobitnú VLAN)

keďže poskytovatelia konektivity sú firmy, ktoré toto robia pre tisíce klientov, pochybujem že je chyba na ich strane. navyše mám rovnaký problém v dvoch rôznych datacentrách a cez rôznych poskytovateľov konektivity (virtual cable providers).