Recently one our 7606-S devices was unexpectly shutdown due to loss of power. When this device returned online, all the OSPF neighbor adj started flapping. Our debugs shows that the OSPF process is establishing a Full neighbor state with each neighbor and suddenly receives an invalid OSPF sequence number and begins the exchange process again. This device has 5 OSPF neighbors, each experiencing the same problem.
After parsing through OSPF adj debugs on the neighbor routers is appears that the neighbors are receiving a blank DBD from the faulty device (almost as if its sending an initial DBD packet) and do not see themselves in the hello packet, so they restart their OSPF process which is triggering the out of sequence packet on the faulty device, causing this never ending cycle.
Initially we cleared the OSPF process and then powered down the box, power back on. Same symptoms on both steps.
We eventually came across a forum post that mentioned a unrelated problem but similar symptom that said to reduce the default timers and surprisingly enough that seems to have been the work around in the meantime, however I see the invalid packet coming through the debugs based on the default timers. Very strange.
I'm open to suggestions for troubleshooting, thanks for any support. Unfortuantely I'm unable to open a request with TAC because our SmartNet contract expired and is under a renewal process so it'll be another few weeks I'm told.
We are running IOS 12.2(33)SRD2