OSPF Part 1

OSPF Part 1

OSPF Summarization

There are two types of summarizations in OSPF

• Summarization from one area to another use the "area range"command
• Summarization of external prefixes use the "summary-address" command

Things to remember about OSPF Summarization

• One of the main reasons to summarize is to reduce the convergence of the domain.

If you are advertising an area prefix into a backbone router in your OSPF network there is no advertisement of this change into the backbone. Meaning there are no LSA advertisements that will be sent out including a network summary LSA.

OSPF Adjacencies

• OSPF adjacencies occur through the exchange of hello packets. Once an adjacency is formed the link state data base is synchronized.

• On point to point links the neighbors become fully adjacent.

• But in the LAN environment all the routers form an adjacency with the DR or designated router, as well as the BDR or backup designated router.

• The DROTHERS maintain a partial relationship with eachother but whenever there is a change all routers in the LAN are updated by the DR.

OSPF Network Types

There are 3 main OSPF network types

1. P2P aka Point-to-Point:In a P2P OSPF network there is no DR or BDR. OSPF will automatically detect a P2P link so there is no need to configure priority values or anything of the sort related to the DR/BDR.
   
2. Broadcast: Broadcast OSPF networks are the most common you will see. Broadcast OSPF networks do have the DR/BDR and there is an election process. Remember that the DR is responsible for forming the OSPF network and informing its neighbors of any changes. The DR election process works like this.The router with the highest IP address or Priority value wins the DR election. The default priority of 1 is given to every router in the network. If you want to manually edit who becomes the DR you need to change the priority value to higher than 1, or if you want to make sure a router never becomes the DR you set its priority to 0.
3. NBMA aka Non Broadcast Multi Access:NBMA is only seen on a frame relay network and is thus becoming legacy but it is still important to know. In an NBMA network the DR/BDR still exist but you have to manually form neighbors using the "neighbor" command. This is pretty much the same as in EIGRP and RIP.

The other network types in OSPF are:

1. P2MP aka Point to Multi Point:P2MP network types do not have a DR/BDR at all, but you do not have to manually configure its neighbors either. OSPF will automatically detect the p2p link as mentioned before.
2. P2MP Non-Broadcast:P2MP Non-Broadcast network types also do not have a DR/BDR because of the p2p link. But you do have to manually configure the neighbors using the "neighbor" command to form an adjacency.
3. Virtual Links:As Anthony Sequiera likes to put it are "temporary band aids" for a network. Meaning they should never be a permanent solution and should only be used to connect an area back to area 0 in case of loss of connectivity.
4. Loopback:Look back network types advertise the /32 bit host route in the network. This can be configured not to advertise the /32 bit host route for security purposes.

Important faqs:

1) There is a DR in an NBMA network type
2) The default priority value is 1
3) Setting a higher priority on a device does not preempt the existing DR.

OSPF Graceful Restart

OSPF Graceful Restart is a killer technology that allows a router with multiple processors to restart the control plane without shutting down the entire router so the neighbors do not recalculate the topology as if the router went down. OSPF graceful restart is not limited to OSPF but is also available in IS-IS, EIGRP, LDP, and BGP. OSPF graceful restart is aka NSF or NSR depending on what cisco press book you are reading.

How the process works is first off the router has to have 2 processors to make this happen. One processor will restart the control plane while the other processor continues to forward traffic aka hardware based forwarding while restarting the control plain of the second processor. During the graceful restart messages will be sent to all neighbors which all routers will recognize that explains a graceful restart is happening and that the DR is not actually dead. There is a special keep alive message that I am not familiar with that informs the routers not to recalculate the topology. This stabilizes the peer devices as long as they understand the graceful restart messages. Most older platforms, and older ios's will not understand this message.

Important FAQS:

• The hardware base forwarding capability forwards traffic using a secondary processor during an OSPF graceful restart.

OSPF Troubleshooting (very important)

When troubleshooting an empty OSPF Neighbor List the common causes of this problem are listed below:

• Not properly configured on interfaces
• Layer 1 or 2 issue
• Passive interface
• Access list blocking
• Error in Layer 3 interface configuration
• Hello or Dead timer mismatch
• Authentication configuration error
• Area ID mismatch
• Stub flag mismatch
• Secondary IP addressing issue
• Incorrect network type configuration

When troubleshooting a "Stuck in Attempt State" the common causes of this problem are listed below:

• Misconfigured neighbor statement
• Unicast non-functional in NBMA environment

When troubleshooting a "Stuck in INIT State" the common causes of this problem are listed below:

• Hellos being blocked in one direction
• Multicast nonfunctional on one side
• Authentication only on one side
• Broadcast keyword missing from map command

When troubleshooting a "Stuck in Two-WAY State" message the common causes of this problem are listed below:

• Priority of 0 on all routers
• Neighbor stuck in EXSTART
• Mismatched MTU
• Duplicate Router ID's
• Broken unicast connectivity

When troubleshooting a "Stuck in LOADING State" the common causes of this problem are listed below:

• Mismatched MTU
• Corrupted link-state request