Monitoring Troubleshooting an OSPF Network | OSPF Network Design Solutions

Table of Contents

The following list provides explanations for representative lines from the preceding output where a newly created entry has been added to the IP routing table.

• metric change. Indicates that this entry existed previously, but its metric changed and the change was reported by means of OSPF. The numbers inside the brackets report the administrative distance and the actual metric.

• cache invalidation. Indicates that the fast-switching cache was invalidated due to a routing table change.

• new version. Indicates the version number of the routing table. When the routing table changes, this number is incremented. The hexadecimal numbers are internal numbers that vary from version to version and software load to software load.

In the following output, the holddown and cache invalidation lines are displayed. Most of the distance vector routing protocols use holddown to avoid typical problems like counting to infinity and routing loops. If you look at the output of show ip protocols, you will see what the timer values are for holddown and cache invalidation.

cache invalidation corresponds to came out of holddown. delete route is triggered when a better path comes along. It gets rid of the old inferior route.

    RT: delete route to 172.26.219.0 via 172.24.76.30, igrp metric    [100/10816]    RT: no routes to 172.26.219.0, entering holddown    IP: cache invalidation from 0x115248 0x1378A, new version 5737    RT: 172.26.219.0 came out of holddown

Simple IP Troubleshooting Tools

Everyone is aware that networks break and must be repaired as quickly as possible. However, what no one is aware of is that you can isolate and correct a variety of network problems without expensive tools. This section discusses two of the more common features available to network engineers, ping and trace route. These two tools are used specifically for diagnosing IP connectivity within a network.

Pings and Extended Pings

Ping is a common network diagnostic tool found in UNIX stations, Windows NT, Windows 95, and your Cisco routers. It might help you uncover the cause for the various network difficulties. In an IP-based network (such as one running OSPF), the ping command is very useful. Ping sends a burst of data to the designated target and awaits a response.

What Is a Ping?

Ping is a network diagnostic tool that is included in all TCP/IP protocol stacks and is commonly used to verify connectivity to a particular system on a network. Ping sends an Internet Control Messaging Protocol or ICMP (type 8) “echo request” in the form of a data packet to a remote host and displays the results for each returned packet known as a (type 0) “echo reply.” The ping command also displays the time for a response to arrive in milliseconds (minimum/average/maximum).

When performing a ping, you are also verifying that IP is working and that you have connectivity to the site. This can help you determine protocol-related problems.

General Characteristics

The ping feature can be administratively disabled if needed. Disabling ping is typically done to preserve or enhance security within a network. The default ping packet size is 100 bytes within a Cisco router.

A successful ping will appear as follows:

    OSPF_Router# ping 10.251.11.1    Type escape sequence to abort.    Sending 5, 100-byte ICMP Echoes to 10.251.11.1, timeout is 2 seconds:    !!!!!    Success rate is 100 percent (5/5), round-trip min/avg/max = 4/4/4 ms

An unsuccessful ping will appear as follows:

    OSPF_Router# ping 10.251.10.1    Type escape sequence to abort.    Sending 5, 100-byte ICMP Echoes to 10.251.10.1, timeout is 2 seconds:    .....    Success rate is 0 percent (0/5)

Are Packets Being Dropped?

Ping inserts a sequence number in each packet it sends and notes which sequence number it receives back and in what order. This can be used to determine if packets are being dropped within your network or if they are being reordered. Both problems will require further investigation.

Is the Network Corrupting Your Packets?

The ping packet also contains the standard IP checksum. If the packet is corrupted or damaged, then the checksum will reflect it.

Round Trip Time (RTT)

Each packet also receives a timestamp when it is sent. This timestamp is then echoed back and it will be displayed for a response to arrive in milliseconds. There will be three values available to you: minimum, average, and maximum.

To verify your findings, determine what the expected round trip time should be. This can be calculated as follows:

One Way Trip: (packet size) * 8 / bandwidth in Kbps

For example, on a 56Kbps circuit with a 512 byte packet, you will calculate the round trip time as 512 * 8 / 56000 = .073 or 73 ms. If you want round trip, you will multiply 73 ms by 2 for a result of 146 ms as the average round trip time.

Please note that calculating the RTT can sometimes vary. This can occur on packet switched networks (Frame Relay), bandwidth changes, traffic congestion or utilization. A good rule of thumb is that if the RTT briefly increases or becomes sporadic, then do not be immediately concerned. Though steady increases over time would warrant further investigation.

TIPS:
A router will typically place a ping in a lower priority than customer traffic, which can add 10ms to the round trip time.

Stress Testing

Ping will also allow you to perform a stress test. Obviously the normal packet size of 64 bytes will not suffice. In this case, you will want to use an extended ping, as described in the section that follows, to increase the packet size. You will first want to record the statistics on the interface you are stress testing. Then send a very large number of 1,500-byte packets (a minimum of 10,000 packets is the recommended standard).

Table of Contents