Task 2.1 - IOS Behavioral Issue??

I was running IOS 12.4(24)T8 on all of my IOS devices (thinking newest is best); but, after watching INE's advanced troubleshooting vSeminar, Brian Dennis specifically states about 5-10 minutes into the intro video to run 12.4(15)T on all routers and 12.2(44)SE on all of the 3560's (running 5 1841's, 1 3825, 4 3560's). 

Anyway, on task 2.1, the OSPF peerings will never come up if you don't adjust the network type on the frame-relay to PTP on R3/4 and PTMP on R5.  If you don't do that, it never sends the hello packet on the frame relay links to form the adjacency.

I thought that maybe a reboot would've fixed it, but that didn't seem to work either.  What's even more interesting is that prior to the reboot (applying the beginning of the config after a 'write erase') the OSPF network type was "Broadcast"; however, after a 'wr mem' and a reboot, the OSPF network type was "NON_BROADCAST" - obviously requiring a network statement.

Has anybody else run into this?


  • JoeMJoeM ✭✭✭

    Hi EMelloul,

    This debug will show the neighbor attempts.

                     debug ip ospf adjacency

    For this lab, the  Hub and Spokes (R3-R5) are "broadcast".  It is not necessary to change it.

    Did you set the priority to "0" on the spokes?  If not, this will make things random  DR/BDR/DROTHER)


    Write Mem  destroyed your initial config completely.  This is why, after rebooting, your interfaces came up with the default network type - nonbroadcast.  This would require neighbor commands. Edit:  I read that as write erase.   Going for morning coffee. ;-)

    Be cautious with mixing the network types.  Neighbors should match network types exactly.  If you want to make the hub P2MP, do the same for the spokes.   With that said, this lab does work with the initial settings (broadcast).    OSPF
    network types are very important. I am confident that your OS version would not change the behavior for this.


    Cisco Doc:   Typical Reasons for OSPF Neighbor Problems

Sign In or Register to comment.