OSPF bounces
Weird Problem between two Cisco devices.
On one device we are seeing errors in the log (router A)
%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on VlanY from EXSTART to DOWN, Neighbor Down: Too many DBD retransmitions
%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on VlanY from DOWN to DOWN, Neighbor Down: Dead timer expired
and on router B the head end router
Now at first look this would seem to be a physical problem of some sort, we are not able to see any kind of errors, So on some investigation we discovered that the ARP table is not getting updated on the one side of the link. And the sniffer shows (from either side) the pings between the links source with the same IP address no matter where the traffic is coming from??
Then the really weird part is this is happening on two different sets of routers in the same location. (the other half of the load for each).
About a month prior to this happening we bounce the interfaces and they both started to work.
but only for a few hours then both quit about the same time.
we tried that again last night to no avail. Cisco TAC has had us trying a few other things and none of them made a difference. now each router has the redundancy for the other vlan that is affect and is working normal.
Next option among the best guess for solution is to replace the Gbic at the head end side, that is the side that does not have the arp entry.
Later is was identified that 3 mac addresses matching the other switch were causing the issue. Both switches are trying to use the same MAC address and this is causing contention.
On one device we are seeing errors in the log (router A)
%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on VlanY from EXSTART to DOWN, Neighbor Down: Too many DBD retransmitions
%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on VlanY from DOWN to DOWN, Neighbor Down: Dead timer expired
and on router B the head end router
| %DTP-SP-5-DOMAINMISMATCH: Unable to perform trunk negotiation on port Gi5/1 because of VTP domain mismatch. | |
| %EARL_L3_ASIC-SP-3-INTR_WARN: EARL L3 ASIC: Non-fatal interrupt Packet Parser block interrupt | |
| %MLS_STAT-SP-4-IP_LEN_ERR: MAC/IP length inconsistencies | |
Now at first look this would seem to be a physical problem of some sort, we are not able to see any kind of errors, So on some investigation we discovered that the ARP table is not getting updated on the one side of the link. And the sniffer shows (from either side) the pings between the links source with the same IP address no matter where the traffic is coming from??
Then the really weird part is this is happening on two different sets of routers in the same location. (the other half of the load for each).
About a month prior to this happening we bounce the interfaces and they both started to work.
but only for a few hours then both quit about the same time.
we tried that again last night to no avail. Cisco TAC has had us trying a few other things and none of them made a difference. now each router has the redundancy for the other vlan that is affect and is working normal.
Next option among the best guess for solution is to replace the Gbic at the head end side, that is the side that does not have the arp entry.
Later is was identified that 3 mac addresses matching the other switch were causing the issue. Both switches are trying to use the same MAC address and this is causing contention.
