On Mon, Sep 11, 2017 at 18:41 -0400, sven falempin wrote: > Feels like it s impossible to use virtual routing table without a rdomain > on interface with 6.1 >
I think you were relying on an arp(1) bug that influenced your understanding of routing tables and routing domains. To recap: 1. There can be only one ARP table per routing domain. > # arp -V 122 -s 172.16.1.1 ac:64:dd:b0:00:03 [permanent] > arp: writing to routing socket: No such process > arp: 172.16.1.1: No such process > What you're trying to do here is to add an ARP entry on the routing domain 122 (not routing table 122). 2. A single routing domain can have multiple routing tables. > Even if the routing can be modify with > > /sbin/route -T122 add -inet 172.16.1.1 -llinfo -link -static -iface em5 && > route -T122 add default 172.16.1.1 > This adds a default route into the routing table number 122 no matter which routing domain it belongs to (0 in your case). > - > > # route -T122 show > Routing tables > > Internet: > Destination Gateway Flags Refs Use Mtu Prio > Iface > default 172.16.1.1 UGS 0 0 - 8 em5 > 172.16.1.1 link#6 UHLS 0 1 - 8 em5 > > - > > arp table remain empty and i cannot ping -V 122 172.16.1.1 > > This was possible 6.0 > > Will it be possible in 6.2 ?? YES :-D :o > arp(1) had a bug that prevented all of it's execution modes from working on multiple rdomains, so it appears that you were relying on "arp -V 122" working on rdomain 0 while in fact the intention was for it work work on rdomain 122 and this should be the case now. Hope this clarifies the situation. > # uname -a > OpenBSD lulz.com 6.2 GENERIC.MP#63 amd64 > > # ifconfig em0 > em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 > lladdr 00:03:2d:36:95:5e > index 1 priority 0 llprio 3 > groups: egress > media: Ethernet autoselect (1000baseT full-duplex,rxpause,txpause) > status: active > inet 172.16.1.45 netmask 0xffffff00 broadcast 172.16.1.255 > > # route -T 120 add default 172.16.1.1 > add net default: gateway 172.16.1.1 > # ping -V 172.16.1.1 > ping: rtable value is invalid: 172.16.1.1 > # ping -V 120 172.16.1.1 > PING 172.16.1.1 (172.16.1.1): 56 data bytes > 64 bytes from 172.16.1.1: icmp_seq=0 ttl=255 time=0.444 ms > 64 bytes from 172.16.1.1: icmp_seq=1 ttl=255 time=0.351 ms > > > -- > > I m asking a lot but > Is there a chance to fix this in 6.1 without to much work ?? > Any workaround ( better than mine ) ? > > [0]-[]-[/] > # ifconfig em5 > em5: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 > lladdr fe:e1:ba:28:22:44 > description: citywan > index 6 priority 0 llprio 3 > groups: egress > media: Ethernet autoselect (100baseTX full-duplex,rxpause,txpause) > status: active > inet 172.16.1.33 netmask 0xffffff00 broadcast 172.16.1.255 > [0]-[]-[/] > # route -T 111 add default 172.16.1.1 > add net default: gateway 172.16.1.1: Network is unreachable > You should send this bug report separately. Do I understand it correctly that this command doesn't work util you issue the one below? /sbin/route -T122 add -inet 172.16.1.1 -llinfo -link -static -iface em5 > i am not used to these king of behavior change with openbsd. > > Best. > > -- > -- > --------------------------------------------------------------------------------------------------------------------- > Knowing is not enough; we must apply. Willing is not enough; we must do
