ospf redistribute connected subnets ... not working?

[Chris Schroeder]

First off, I am just trying to figure out this problem that I inherited. ;-)

The environment:  

Hardware: 6500 w/sup720

One active routing protocol:  OSPF 

One active area (100)

Problem: No subinterfaces under physical interface 4/2 show up in the OSPF routing table even though the "redistribute connected subnets" command is used under OSPF.

Interface configuration

-------------------------------------------

interface GigabitEthernet4/2
no ip address
!
interface GigabitEthernet4/2.600
encapsulation dot1Q 600
ip address 10.1.111.250 255.255.255.0
standby 0 ip 10.1.111.254
standby 0 priority 125
standby 0 preempt
!
interface GigabitEthernet4/2.800
encapsulation dot1Q 800
ip address 10.1.110.250 255.255.255.0
standby 0 ip 10.1.110.254
standby 0 priority 125
standby 0 preempt
!
interface GigabitEthernet4/2.802
encapsulation dot1Q 802
ip address 10.1.112.250 255.255.255.0
standby 0 ip 10.1.112.254
standby 0 priority 125
standby 0 preempt
!

-------------------------------------------

OSPF configuration (partially truncated in interfaces but no interface 4/2 or subinterface of 4/2 is included under the no passive-interface commands because they're directly connected and the "redistribute connected subnets" should pick them up, right?)

-------------------------------------------

router ospf 100
router-id 10.1.120.47
log-adjacency-changes
redistribute connected subnets
passive-interface default
no passive-interface GigabitEthernet3/2
no passive-interface GigabitEthernet3/3
no passive-interface GigabitEthernet3/4
no passive-interface GigabitEthernet3/5

...etc...
no passive-interface Vlan4001
network 10.0.0.0 0.255.255.255 area 0
!

-------------------------------------------

relevant configuration:

router# show ip route 10.1.110.1
Routing entry for 10.1.110.0/24
Known via "connected", distance 0, metric 0 (connected, via interface)
Redistributing via ospf 100
Routing Descriptor Blocks:
* directly connected, via GigabitEthernet4/2.800
Route metric is 0, traffic share count is 1

router# show ip int br | in 4/2.800
GigabitEthernet4/2.800 10.1.110.250 YES NVRAM up up
router#show ip route ospf | in 10.1.110.1
<NO MATCH ON LAST COMMAND - should show up as a 10.1.110.0/24 route>

 -------------------------------------------

Problem: None of the L3 subinterface subnets under interface gi4/2 show up in the ospf routing table at all.  Routes for all the other L3 interfaces on the router show up fine, but they're not configured as subinterfaces, just normal L3 interfaces with an IP.  No simultaneous trunking or anything like that.

Which obvious thing am I overlooking?

[Harold Ritter]

Hi,

It is normal for the "show ip route ospf" comamnd not to show 10.1.110.1 for 2 reasons.

1. subnet 10.1.110.0/24 will be seen as a connected route, not as ospf route on tis router.

2. even if it was know as an ospf route, your include matches on 10.1.110.1 rather than on 10.1.110.0

If you want to see whether these routes are correctly being redistributed in ospf, I would recommend you do a "show ip ospf database external".

Regards,

[Chris Schroeder]

Thanks in advance for the assistance.

"show ip ospf database external" with no include returns lots of information about routes being advertised from this particular 6500 but not the subnet itself, so I can't tell from the this command output whether or not 10.1.110.0/24 or 10.1.111.0/24 are actually being advertised.

[Harold Ritter]

Try the following: show ip ospf database external | include 10.1.110.0

Regards,

[Chris Schroeder]

router#show ip ospf database external | include 10.1.110.0
router#

...no hits.   ...and I see that they're calling the subnets "Link State ID"....(wtf?)  anyway :)   There are a bunch of subnets being advertised, including five from this particular router ID in OSPF, but the three L3 subinterfaces under 4/2 are NOT among them.

partial unfiltered output....

INET-DFW1-RTR-020#show ip ospf database external

OSPF Router with ID (10.1.120.47) (Process ID 100)

Type-5 AS External Link States

Routing Bit Set on this LSA
LS age: 1318
Options: (No TOS-capability, DC)
LS Type: AS External Link
Link State ID: 10.10.33.16 (External Network Number )
Advertising Router: 10.1.120.53
LS Seq Number: 80000BDA
Checksum: 0x9E6B
Length: 36
Network Mask: /28
Metric Type: 2 (Larger than any link state path)
TOS: 0
Metric: 1
Forward Address: 0.0.0.0
External Route Tag: 3489725448

Routing Bit Set on this LSA
LS age: 1318
Options: (No TOS-capability, DC)
LS Type: AS External Link
Link State ID: 10.10.33.32 (External Network Number )
Advertising Router: 10.1.120.53
LS Seq Number: 80000BDA
Checksum: 0xFDFB
Length: 36
Network Mask: /28
Metric Type: 2 (Larger than any link state path)
TOS: 0
Metric: 1
Forward Address: 0.0.0.0
External Route Tag: 3489725448

Routing Bit Set on this LSA
LS age: 1320
Options: (No TOS-capability, DC)
LS Type: AS External Link
Link State ID: 10.10.33.112 (External Network Number )
Advertising Router: 10.1.120.53
LS Seq Number: 80000BDA
Checksum: 0xDACE
Length: 36
Network Mask: /28
Metric Type: 2 (Larger than any link state path)
TOS: 0
Metric: 1
Forward Address: 0.0.0.0
External Route Tag: 3489725448

Routing Bit Set on this LSA
LS age: 1320
Options: (No TOS-capability, DC)
LS Type: AS External Link
Link State ID: 10.10.33.128 (External Network Number )
Advertising Router: 10.1.120.53
LS Seq Number: 80000BDA
Checksum: 0x3A5F
Length: 36
Network Mask: /28
Metric Type: 2 (Larger than any link state path)
TOS: 0
Metric: 1
Forward Address: 0.0.0.0
External Route Tag: 3489725448

....and five of the routes are actually coming from this 6500:

INET-DFW1-RTR-020#show ip ospf database external | in 10.1.120.47
OSPF Router with ID (10.1.120.47) (Process ID 100)
Advertising Router: 10.1.120.47
Advertising Router: 10.1.120.47
Advertising Router: 10.1.120.47
Advertising Router: 10.1.120.47
Advertising Router: 10.1.120.47
INET-DFW1-RTR-020#

.... but none of those are the 10.1.100.0/24 or 10.1.110.0/24 subnets which are directly connected as subinterfaces under gi4/2

[Harold Ritter]

The reason you are not seeing these routes as external might be that they are being included as internal due to your network statement that includes all these sub interfaces. Can you do a "show ip ospf interface brief" to verify that?

Regards,

[Chris Schroeder]

Harold:  Here's the output you wanted.... which seems to indicate that L3 for Vlans 800 and 600 are being advertised... which leaves me with my original problem, I don't understand why communication direct between 800 and 600 isn't working.

router#show ip ospf int br
Interface PID Area IP Address/Mask Cost State Nbrs F/C
Lo0 100 0 10.1.120.47/32 1 LOOP 0/0
Vl4001 100 0 10.6.0.5/30 1 BDR 1/1
Gi4/3 100 0 10.6.0.181/30 100 BDR 1/1
Gi3/46 100 0 10.6.0.170/30 1 DR 1/1
Gi3/2 100 0 10.6.0.53/30 1 DOWN 0/0
Gi3/48 100 0 10.6.0.129/30 1 BDR 1/1
Gi4/48 100 0 10.6.0.133/30 2 BDR 1/1
Gi4/2.802 100 0 10.1.112.250/24 1 DR 0/0
Gi4/2.800 100 0 10.1.110.250/24 1 DR 0/0
Gi4/2.600 100 0 10.1.111.250/24 1 DR 0/0
Gi3/47 100 0 10.6.0.46/30 1 DOWN 0/0
Gi3/5 100 0 10.6.0.94/30 1 DOWN 0/0
Gi3/4 100 0 10.6.0.65/30 1 DOWN 0/0
Gi3/3 100 0 10.6.0.81/30 1 DR 1/1
router#

[pigallo]

Hi,

if you issue the command show ip ospf database adv-router 10.1.120.47 you will see every LSA in the database that is advertised by that node.

Kindly check with this command if you are advertising those routes as internal routes (because of network command) or external routes (because of redis conn subnets command).

[Chris Schroeder]

Harold and Pigallo, I believe the following output will answer both of your questions and it confirms my hypothesis - the directly connected subnets are NOT being redistributed in OSPF.  More on that at the end of this message.

Output:

router#show ip ospf database adv-router 10.1.120.47

OSPF Router with ID (10.1.120.47) (Process ID 100)

Router Link States (Area 0)

Link ID ADV Router Age Seq# Checksum Link count
10.1.120.47 10.1.120.47 1337 0x8000B5D9 0x004A97 10

Net Link States (Area 0)

Link ID ADV Router Age Seq# Checksum
10.6.0.81 10.1.120.47 1337 0x80000022 0x00A901
10.6.0.170 10.1.120.47 827 0x80000963 0x006196

Type-5 AS External Link States

Link ID ADV Router Age Seq# Checksum Tag
50.16.201.61 10.1.120.47 827 0x80003C5C 0x0056B6 0
50.56.174.29 10.1.120.47 827 0x80003C5C 0x00DF40 0
166.78.137.240 10.1.120.47 827 0x80003C5C 0x003FA7 0
166.78.138.23 10.1.120.47 827 0x80003C5C 0x00B609 0
198.101.238.77 10.1.120.47 827 0x80003C5C 0x00915C 0

router#show run | section router ospf
router ospf 100
router-id 10.1.120.47
log-adjacency-changes
redistribute connected subnets
passive-interface default
no passive-interface GigabitEthernet3/2
no passive-interface GigabitEthernet3/3
no passive-interface GigabitEthernet3/4
no passive-interface GigabitEthernet3/5
no passive-interface GigabitEthernet3/46
no passive-interface GigabitEthernet3/47
no passive-interface GigabitEthernet3/48
no passive-interface GigabitEthernet4/1
no passive-interface GigabitEthernet4/3
no passive-interface GigabitEthernet4/48
no passive-interface Vlan4001
network 10.0.0.0 0.255.255.255 area 0

router#show run | begin interface GigabitEthernet4/2
interface GigabitEthernet4/2
no ip address
!
interface GigabitEthernet4/2.600
encapsulation dot1Q 600
ip address 10.1.111.250 255.255.255.0
standby 0 ip 10.1.111.254
standby 0 priority 125
standby 0 preempt
!
interface GigabitEthernet4/2.800
encapsulation dot1Q 800
ip address 10.1.110.250 255.255.255.0
standby 0 ip 10.1.110.254
standby 0 priority 125
standby 0 preempt
!
interface GigabitEthernet4/2.802
encapsulation dot1Q 802
ip address 10.1.112.250 255.255.255.0
standby 0 ip 10.1.112.254
standby 0 priority 125
standby 0 preempt
!

router#show ip int br | in 4/2
GigabitEthernet4/2 unassigned YES NVRAM up up
GigabitEthernet4/2.600 10.1.111.250 YES NVRAM up up
GigabitEthernet4/2.800 10.1.110.250 YES NVRAM up up
GigabitEthernet4/2.802 10.1.112.250 YES NVRAM up up

Now here's the kicker, and why all this is a problem, and why I suspect routing to be the culprit.   This particular 6500 has a direct layer 3 connection to vlans 600 and 800 (and 802).   But look at this output:

First, ping a known good windows host on vlan 800 from this 6500:

router#ping 10.1.110.100

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.1.110.100, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/4 ms

Yay, right?

Ok, now ping the SAME host from this same switch on Vlan800 but using a source IP from Vlan600, which is also on the same switch:

router#ping 10.1.110.100 source Gi4/2.600

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.1.110.100, timeout is 2 seconds:
Packet sent with a source address of 10.1.111.250
.....
Success rate is 0 percent (0/5)

Boo.   Hence, my problem.  

Help!

[Chris Schroeder]

Additional info:

router#show ip ospf nei

Neighbor ID Pri State Dead Time Address Interface
10.1.120.5 1 FULL/DR 00:00:33 10.6.0.6 Vlan4001
10.1.120.56 1 FULL/DR 00:00:34 10.6.0.182 GigabitEthernet4/3
10.1.120.54 1 FULL/BDR 00:00:38 10.6.0.169 GigabitEthernet3/46
10.1.120.48 1 FULL/DR 00:00:37 10.6.0.130 GigabitEthernet3/48
10.1.120.48 1 FULL/DR 00:00:31 10.6.0.134 GigabitEthernet4/48
10.1.120.39 1 FULL/BDR 00:00:32 10.6.0.82 GigabitEthernet3/3
router#

[pigallo]

Hi,


ok from the output we do not see them as redistributed into Ospf.

However since "show ip ospf interface brief" output reveals that they participate into ospf process you should advertise them somehow.

Run this command please: "show ip ospf database router adv-router 10.1.120.47"

you will see all type 1 LSAs advertised by the switch.

I suspect, since the dot 800 and 600 interfaces are still passive interfaces, that they should appear as LSA 1 stub networks in the LSDB.

[Harold Ritter]

Hi,

Hosts on two directly connected subnets should be able to ping each other without the need of a routing protocol as long as they have the proper default gateway configured. I think the issue you are experiencing is not related to OSPF at all. Can you check that the hosts are properly configured.

Regards,