[Xorp-users] OSPF point-to-multipoint issue

Vincent, Michael - 0665 - MITLL michael.vincent at ll.mit.edu
Fri Jun 18 12:21:11 PDT 2010


> Ahh, so to get to the original peer, it needs to first remove the routes on the secondary link?

Not necessarily, you can still get there (data path) through the multi-hop path; however, OSPF won't form an adjacency over a multi hop.  So to that point, yes, the host route that OSPF added needs to be removed for OSPF to then re-form it's adjacency.  Catch-22 sounds appropriate.


I've only included the outputs for nodes 1 and 3 as that is what the example is talking about.  The actual experiment fails other links also and thus more peering sessions go down and do not re-establish.


START:
root at nlet-89-1> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.2       eth1/eth1              Full      10.10.1.2        128     2
172.16.1.3       eth1/eth1              Full      10.10.1.3        128     3
172.16.1.4       eth1/eth1              Full      10.10.1.5        128     3
172.16.1.5       eth1/eth1              Full      10.10.1.5        128     2
172.16.1.6       eth1/eth1              Full      10.10.1.6        128     2
172.17.2.1       eth2.1212/eth2.1212    Full      10.10.1.8          1    31
root at nlet-89-1>

--

[root at nlet-89-3 DISA]# xorpsh -c "show ospf4 neighbor"
Welcome to XORP on nlet-89-3
root at nlet-89-3> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.1       eth1/eth1              Full      10.10.1.1        128     2
172.16.1.2       eth1/eth1              Full      10.10.1.2        128     3
172.16.1.4       eth1/eth1              Full      10.10.1.4        128     2
172.16.1.5       eth1/eth1              Full      10.10.1.5        128     3
172.16.1.6       eth1/eth1              Full      10.10.1.6        128     3
root at nlet-89-3> [root at nlet-89-3 DISA]#


INTERMEDIATE:
root at nlet-89-1> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.2       eth1/eth1              Down      10.10.1.2          0     0
172.16.1.3       eth1/eth1              Full      10.10.1.3        128     3
172.16.1.4       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.5       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.6       eth1/eth1              Down      10.10.1.6          0     0
172.17.2.1       eth2.1212/eth2.1212    Full      10.10.1.8          1    31
root at nlet-89-1>

--

[root at nlet-89-3 DISA]# xorpsh -c "show ospf4 neighbor"
Welcome to XORP on nlet-89-3
root at nlet-89-3> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.1       eth1/eth1              Down      10.10.1.1          0     0
172.16.1.2       eth1/eth1              Down      10.10.1.2          0     0
172.16.1.4       eth1/eth1              Full      10.10.1.4        128     2
172.16.1.5       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.6       eth1/eth1              Down      10.10.1.6          0     0
root at nlet-89-3> [root at nlet-89-3 DISA]#


FINAL:
root at nlet-89-1> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.2       eth1/eth1              Down      10.10.1.2          0     0
172.16.1.3       eth1/eth1              Full      10.10.1.3        128     3
172.16.1.4       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.5       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.6       eth1/eth1              Down      10.10.1.6          0     0
172.17.2.1       eth2.1212/eth2.1212    Full      10.10.1.8          1    31
root at nlet-89-1>

--

[root at nlet-89-3 DISA]# xorpsh -c "show ospf4 neighbor"
Welcome to XORP on nlet-89-3
root at nlet-89-3> show ospf4 neighbor
  Address         Interface             State      ID              Pri  Dead
172.16.1.1       eth1/eth1              Down      10.10.1.1          0     0
172.16.1.2       eth1/eth1              Down      10.10.1.2          0     0
172.16.1.4       eth1/eth1              Full      10.10.1.4        128     2
172.16.1.5       eth1/eth1              Down      10.10.1.5          0     0
172.16.1.6       eth1/eth1              Down      10.10.1.6          0     0
root at nlet-89-3> [root at nlet-89-3 DISA]#



    Cheers.

-- 
Michael J. Vincent
Group 65 - Advanced Networks and Applications
MIT Lincoln Laboratory
244 Wood Street
Lexington, MA 02420
Office:  +1.781.981.3459


-----Original Message-----
From: Ben Greear [mailto:greearb at candelatech.com] 
Sent: Friday, June 18, 2010 2:52 PM
To: Vincent, Michael - 0665 - MITLL
Cc: xorp-users at xorp.org; Zuena, John - 0665 - MITLL
Subject: Re: [Xorp-users] OSPF point-to-multipoint issue

On 06/18/2010 11:43 AM, Vincent, Michael - 0665 - MITLL wrote:
> Yes, show interfaces shows the interface in question (the ones OSPF does peering on) as ENABLED.
>
> Additionally, if we exit from xorpsh and remove the offending host routes from linux (route del x.x.x.x) (the ones Xorp OSPF should be removing when OSPF converges) the OSPF sessions are re-established.  This is further feeding my belief that Xorp is not properly removing OSPF routes during OSPF convergence.

Ahh, so to get to the original peer, it needs to first remove the routes on the secondary link?

Maybe xorp doesn't bother removing routes until it establishes FULL access with the
new router, which would be a neat catch-22 scenario in your case.

Can you post the results of 'show ospf4 neighbor'
for each of your steps:

*  Initial config  (This is working OK, right)
*  First link failure (And this is OK too)
*  Second link failure  ( But this fails to switch back to initial config setup??)

Thanks,
Ben

-- 
Ben Greear <greearb at candelatech.com>
Candela Technologies Inc  http://www.candelatech.com




More information about the Xorp-users mailing list