Hallo!
I am thinking of startig using ospf techology to set up higer redundancy
but at the moment i am just there where i am trying out my first setup.
I have read some books on the topic and now i am following the text from
http://www.openbsd.org/papers/linuxtag06-network.pdf.
To follow the 'Distant-office - Local-office' part from page 7 i set up
xen hvm hosts, configured networking and started ospfd daemons on both
ends, when they start i belive things work as they should but after
taking preferred link down and putting back up my ospf setup isnt
adjusting to the change.
I am sorry for the long letter but i decided to present my setup in
appropriate detail so someone could point me to the right direction.
Leaving out details my setup is like this
ne1, 192.168.15.1 ---------- 192.168.15.254, ne1
metric 100, over 1st xen bridge
distant local
office office
metric 10, over 2nd xen bridge
ne2, 192.168.16.1 ---------- 192.168.16.254, ne2
ospfd.conf at distant office is
router-id 0.0.0.2
redistribute connected
# areas
area 0.0.0.0 {
interface ne1 { metric 100 }
interface ne2 { metric 10 }
}
ospfd.conf at local office is
router-id 0.0.0.1
redistribute connected
# areas
area 0.0.0.0 {
interface ne1 { metric 100 }
interface ne2 { metric 10 }
}
I started ospf like '# ospfd -vd'.
1. FIRST TEST - both links are up and things work
And after ospf settles and things work i have at distant office
(interesting why linkstate is unknown, because of xen?)
# ospfctl show int
Interface Address State HelloTimer Linkstate Uptime nc ac
ne2 192.168.16.1/24 DR 00:00:07 unknown 00:02:33 1 1
ne1 192.168.15.1/24 DR 00:00:07 unknown 00:02:33 1 1
and neighborhood
# ospfctl show nei
ID Pri State DeadTime Address Iface Uptime
0.0.0.1 1 FULL/BCKUP 00:00:38 192.168.16.254 ne2 00:05:12
0.0.0.1 1 FULL/BCKUP 00:00:39 192.168.15.254 ne1 00:05:12
and rib
c# ospfctl show rib
Destination Nexthop Path Type Type Cost Uptime
0.0.0.1 192.168.16.254 Intra-Area Router 10 00:05:47
192.168.15.0/24 192.168.15.1 Intra-Area Network 100 00:05:52
192.168.16.0/24 192.168.16.1 Intra-Area Network 10 00:05:52
0.0.0.0/0 192.168.16.254 Type 1 ext Network 110 00:05:47
2. SECOND TEST - preferred ie with metric 10 link down
Things work, at distant office i have
# ospfctl show int
Interface Address State HelloTimer Linkstate Uptime nc ac
ne2 192.168.16.1/24 DR 00:00:06 unknown 00:15:45 1 0
ne1 192.168.15.1/24 DR 00:00:06 unknown 00:15:45 1 1
# ospfctl show nei
ID Pri State DeadTime Address Iface Uptime
0.0.0.1 1 DOWN/OTHER 00:01:29 192.168.16.254 ne2 -
0.0.0.1 1 FULL/BCKUP 00:00:32 192.168.15.254 ne1 00:15:49
as i see, distant office learned new default route, very good
# ospfctl show rib
Destination Nexthop Path Type Type Cost Uptime
0.0.0.1 192.168.15.254 Intra-Area Router 100 00:02:12
192.168.15.0/24 192.168.15.1 Intra-Area Network 100 00:16:33
192.168.16.0/24 192.168.15.254 Intra-Area Network 110 00:02:12
0.0.0.0/0 192.168.15.254 Type 1 ext Network 200 00:02:12
3. THIRD TEST - preferred link is put back up
Traffic flows but it isnt switching back to the preferred metric 10
link, at the distant office i have
# ospfctl show int
Interface Address State HelloTimer Linkstate Uptime nc ac
ne2 192.168.16.1/24 DR 00:00:04 unknown 00:22:17 1 1
ne1 192.168.15.1/24 DR 00:00:04 unknown 00:22:17 1 1
for some reason state stays on FULL/OTHER
# ospfctl show nei
ID Pri State DeadTime Address Iface Uptime
0.0.0.1 1 FULL/OTHER 00:00:36 192.168.16.254 ne2 00:00:13
0.0.0.1 1 FULL/BCKUP 00:00:36 192.168.15.254 ne1 00:21:45
and routing stays on metric 100 link
# ospfctl show rib
Destination Nexthop Path Type Type Cost Uptime
0.0.0.1 192.168.15.254 Intra-Area Router 100 00:08:49
192.168.15.0/24 192.168.15.1 Intra-Area Network 100 00:23:10
192.168.16.0/24 192.168.16.1 Intra-Area Network 10 00:01:38
0.0.0.0/0 192.168.15.254 Type 1 ext Network 200 00:08:49
If i continue and take down the metric 10 link, then traffic stops and i get
# ospfctl show nei
ID Pri State DeadTime Address Iface Uptime
0.0.0.1 1 FULL/OTHER 00:00:37 192.168.16.254 ne2 00:07:52
0.0.0.1 1 DOWN/OTHER 00:00:03 192.168.15.254 ne1 -
and when i put it back i dont get traffic and i have
# ospfctl show nei
ID Pri State DeadTime Address Iface Uptime
0.0.0.1 1 FULL/OTHER 00:00:33 192.168.16.254 ne2 00:09:37
0.0.0.1 1 FULL/OTHER 00:00:33 192.168.15.254 ne1 00:00:47
I can also confirm that i can ping over both links the other side, in
those routers is pf disabled and looking at the traffic i see Hello
messages being sent
# tcpdump -ni ne2
tcpdump: listening on ne2, link-type EN10MB
17:59:24.903057 192.168.16.1 > 224.0.0.5: OSPFv2-hello 48: rtrid 0.0.0.2
backbone dr 192.168.16.1 [tos 0xc0] [ttl 1]
17:59:24.903388 192.168.16.254 > 224.0.0.5: OSPFv2-hello 48: rtrid
0.0.0.1 backbone dr 192.168.16.254 [tos 0xc0] [ttl 1]
17:59:34.913201 192.168.16.1 > 224.0.0.5: OSPFv2-hello 48: rtrid 0.0.0.2
backbone dr 192.168.16.1 [tos 0xc0] [ttl 1]
17:59:34.913442 192.168.16.254 > 224.0.0.5: OSPFv2-hello 48: rtrid
0.0.0.1 backbone dr 192.168.16.254 [tos 0xc0] [ttl 1]
I would be very thankful i somebody could take me forward from here.
Best regards,
Imre
PS During these tests the local office displayed similar output.
