Troubleshooting of OSPF Route Selection

OSPF route selection and troubleshooting case description this case involves the LSA selection problem in specific circumstances of OSPF. This problem is related to the FA address and some NSSA features. By analyzing faults one by one, we also summarize relevant knowledge. The following is a real case. To protect customer information, the router output information has been modified. Fault Diagnosis steps R1, R2 and R3 are NSSA regions, and R1 and R2 are backbone regions. The redistribution network 10.10.119.0/27 on R3 is in NSSA. The customer found that R2 learned The NSSA route normally from R3, but R1 learned the E1 route from R2. 1. Collect information and confirm the problem. We found that both R1 and R2 have TYPE 7 and TYPE 5 routes, but R1 chose TYPE5 and R2 chose NSSA. Why? [1] R1 # show ospf database external 10.10.119.0 OSPF Router with ID (10.10.191.131) (Process ID 18181) Type-5 AS External Link States Routing Bit Set on this lsa ls age: 652 Options: (No TOS-capability, DC) LS Type: AS External Link State ID: 10.10.119.0 (External Network Number) Advertising Router: 10.10.191.132 LS Seq Number: 80000399 Checksum: 0x3768 Length: 36 Network Mask:/27 Metric Type: 1 (Comparable directly to link state metric) TOS: 0 Metric: 1 Forward Address: 10.10.72.89 External Route Tag: 1R1 # sh ospf database nssa-external 10.119.0 OSPF Router with ID (10.10.191.131) (Process ID 18181) Type-7 AS External Link States (Area 7) LS age: 312 Options: (No TOS-capability, Type 7/5 translation, DC) LS Type: AS External Link State ID: 10.10.119.0 (External Network Number) Dvertising Router: 10.10.72.89 LS Seq Number: 8000fe93 Checksum: 0x240 Length: 36 Network Mask:/27 Metric Type: 1 (Comparable directly to link state metric) TOS: 0 Metric: 1 Forward Address: 10.10.72.89 External Route Tag: 12. based on the database information, the E1 route learned by R1 should be sent to R1 after R2 converts TYPE7 to type5. Check the route table carefully and find that, although R1 learned the route from R2, the next hop still passes through R3. Why? [2] These ports are all in NSSA. Will there be five types of routes in NSSA? [3] R2 # show route 10.10.119.0 de Tag 1, type NSSA extern 1 10.10.183.142, from 10.10.72.89, via tengie0/0/0/0 10.10.183.98, from 10.10.72.89, via TenGigE0/1/2/0R1 # show route10.119.0 de Tag 1, type extern 1 10.10.183.74, from 10.10.191.132, via TenGigE0/2/0/0 10.10.183.138, from 10.10.191.132, via TenGigE0/0/0/0R1 # show ospf inter ten0/0/0 | I Area Internet Address 10.10.183.73/30, Area 7R1 # R1 # s How ospf inter ten0/0/0 | I Area Internet Address 10.183.20./ 30, Area 73. first, R1 and R2 are ABR, so even if there are 5 types of routes, it is OK. Why did the Routes learned from R2 go from R3 instead? [4] For the FA address, I simply recall that o is unique to OSPF, it is mainly used to tell the vro in the domain not to find ASBR when accessing a route outside of the domain in a specific situation. To find the FA forwarding address, o must only meet specific conditions, the FA address is not "0.0.0.0". For detailed rules, see related documents. If you filter out the FA address, therefore, this external route will not be added to the route table. Only when the FA address is (intra-area) o or IA (inter-area) can this E2 route be placed in the route table. Careful redistribution of static data to OSPF: if the next hop is an address and is directly connected, it will not be added to the route table. On the contrary, the write port is acceptable because it does not comply with the rules, FA all zero o is different from normal region in NSSA. According to "RFC 3101, specially section 2.3 Type-7 LSAs", there must be a FA address during conversion, so if static redistribution is performed, if the next hop is written as a port, FA will become completely zero, which may cause problems. "6. those Type-7 LSAs that are to be translated into Type-5 LSAs must have their forwarding address set. "4. according to the above information, it is clear that although routes are learned from R2, the forwarding still goes out from the port directly connected to R3 By FA. 5. Why does R1 learn E1 from R2 instead of R1? [5] It seems that the problem can be easily reproduced after an environment test. Is there any documentation? [6] 6. I checked the RFC and found the following information to explain the above question: o for why R1 learned E1 routing from R2, according to RFC 2328: "If two routers, both reachable from one another, originate functionally equivalent AS-external-LSAs (I. e ., same destination, cost and non-zero forwarding address), then the LSA originated by the router having the highest OSPF Router ID is used. "Because R2 (132) is better than R1 (131), R2 is converted and sent to R1. O For why R1 chooses E1 instead of NSSA, according to RFC 1587: "When a type-5 LSA and a type-7 LSA are found to have the same type and an equal distance, the following priorities apply (listed from highest to lowest) for breaking the tie. a. any type 5 LSA. b. A type-7 LSA with the P-bit set and the forwarding address non-zero.c. any other type-7 LSA. "Answer summary [1] According to RFC1587, when type-5 and type-7 have the same type and distance, type-5> type-7 (P-bit + FA)> type -7 [2] When LSA has a FA address, it must be forwarded to the FA address instead of to the ASBR. [3] If a vro belongs to NSSA, there will be no type-5 LSA, but it is possible if this vroabr Is An ABR. [4] FA address. [5] According to RFC 2328, if the two routers can reach each other, when the two devices generate an equivalent LSA, the LSA with a high Router ID will be used. [6] RFC 2328 and RFC1587 commands show ospf database external x. x. x. xshow ospf database nssa x. x. x. xshow route x. x. x. x detailshow ospf inter xxx | I xxx