Re: [Linux-vrf-general] Re: [linux-vrf-general]EACH VRF CAN HAVE ITS OWN ZEBRA D
Status: Beta
Brought to you by:
jleu
Menu
▾
▴
Re: [Linux-vrf-general] Re: [linux-vrf-general]EACH VRF CAN HAVE ITS OWN ZEBRA D
|
From: Cho K. H. <dru...@ho...> - 2006-04-13 01:45:53
|
First of all, i don't see anything done wrong.. try following procedures and find a problem.. i) check if each zebra has its own correct interface list. it can be done by "show interface" on the vtysh. for each zebra, there must be at least two interfaces shown. ii) check if each ripd has its own correct interface list. it can be done by "show ip rip interface". ripd receives the interface information from zebra. If "show ip rip interface" reveals nothing, zebra and ripd have not been connected correctly. iii) do tcpdump on any interface(eth0, eth1, eth2 or eth3) to see if any rip packet is passing. this can be done at the linux shell with the following command: tcpdump -i eth0 iv) if everything is working okay, then try "ip route X.X.X.X Y.Y.Y.Y" on the config node of vtysh. out of vtysh, you must be able to see the routing entry you set on the linux shell with the following command "ip route". if it doesn't show what you have added, this is the problem between zebra and the linux kernel which uses netlink for route synchronization. ========================== 행복한 현재가 좋다! dru...@ho... (조경현) >From: �댁�� <jeo...@co...> >To: <jl...@mi...> >CC: <lin...@li...> >Subject: Re: [Linux-vrf-general] Re: [linux-vrf-general]EACH VRF CAN HAVE ITS OWN ZEBRA DAEMON? >Date: Thu, 13 Apr 2006 03:44:09 +0900 > >I tried to run simple ripd for each vrf. >The result is negative. > >Firsta of all, let me give you what I have done first. >And then, let me ask your help. > >1. My network topology > > 134.75.1.0/24 134.77.1.0/24 134.79.1.0/24 > | | | > | | | > --------eth0 v eth0------eth2 v eth0------eth2 v eth0-------- > | HOST A |---------| vrf1 |-------------| vrf1 |---------| HOST C | > --------151 150 150 151 150 151-------- > | VR1 | | VR2 | > --------eth0 eth1 eth3 eth1 eth3 eth0-------- > | HOST B |---------| vrf2 |-------------| vrf2 |---------| HOST D | > --------151 v 150------150 v 151------150 v 151-------- > | | | > | | | > 134.76.1.0/24 134.78.1.0/24 134.80.1.0/24 > >2. Router configuration > 2.1 VR 1 > A. create vrf 1 and assign eth0 and eth2 to vrf 1 > create vrf 2 and assign eth1 and eth3 to vrf 2 > B. assign zebra to each vrf > chvrf 1 zebra -d > chvrf 2 zebra -d --pid_file=# > * I configured zebra with --enable-tcp-zebra option > * I could access zebra vty both from HOST A and HOST B > > C. create two ripd configuration file > 1) ripd_vrf1.conf > ... > router rip > network 134.75.1.0/24 > network 134.77.1.0/24 > networh eth2 > ... > 2) ripd_vrf2.conf > ... > router rip > network 134.76.1.0/24 > network 134.78.1.0/24 > networh eth3 > ... > D. assign ripd to each vrf > chvrf 1 ripd -d --config_file=ripd_vrf1.conf > chvrf 2 ripd -d --config_file=ripd_vrf2.conf --pid_file=# > > 2.1 VR 2 > A. create vrf 1 and assign eth0 and eth2 to vrf 1 > create vrf 2 and assign eth1 and eth3 to vrf 2 > B. assign zebra to each vrf > chvrf 1 zebra -d > chvrf 2 zebra -d --pid_file=# > * I configured zebra with --enable-tcp-zebra option > * I could access zebra vty both from HOST C and HOST D > > C. create two ripd configuration file > 1) ripd_vrf1.conf > ... > router rip > network 134.77.1.0/24 > network 134.79.1.0/24 > networh eth0 > ... > 2) ripd_vrf2.conf > ... > router rip > network 134.78.1.0/24 > network 134.80.1.0/24 > networh eth1 > ... > D. assign ripd to each vrf > chvrf 1 ripd -d --config_file=ripd_vrf1.conf > chvrf 2 ripd -d --config_file=ripd_vrf2.conf --pid_file=# > >3. Result: negative > * Ping does not reach from Host A to Host C, neither from Host B to Host D. > * I checked "show ip route" on zebra vty, but there was no routing table > update. > * By mannual routing table configuration on each vrf on each VR, I got ping > response from Host A to Host C, vice versa. Of coure, I succeeded between > Host B and D. > >Did I miss something? >Did I do something wrong? >Any advice to go through this problem? > >- jeong ho. |
