Diagnosing IP Networks › How to Trace a TCP/IP Route

How to Trace a TCP/IP Route

It is possible to obtain a hop-by-hop record of the route taken by a packet through a network, starting from the source host and finishing at the destination host.

You can perform a traceroute as follows:

• Directly from the TCP/IP : Network Diagnosis Functions menu
• Use the traceroute action against a host on a list or display; for example on a connection list or the IP Node Monitor
• Issue the TRACEROUTE command from the command line

Perform a Traceroute From the Network Diagnosis Functions Menu

A traceroute attempts to trace the route from the issuing point through all hosts and routers to the destination host; otherwise it traces the route to the point at which a break in communication has occurred.

To trace a route to a remote host

1. Enter /IPDIAG at the prompt.

   The TCP/IP : Network Diagnosis Functions menu appears.

2. Select TR - Trace Route and complete the following field:

   Host Name/Addr

   Specifies the name or IP address of the remote host you want to access.

   If you want to use a specific address of the local stack interface as the issuing point, select it in the Source Address field. If you want to trace the route through a specific physical interface, select it in the Interface field.

   Press Enter.

   The TCP/IP : Trace Route Result List panel appears.

Use Traceroute Action

The Trace Route Result List panel displays the results of the traceroute action. This display is progressively updated as information arrives.

The most likely reason for performing a traceroute action is that a ping is unable to reach a remote host or it is indicating slow response times in the TCP/IP network.

If the host is unreachable, you can take the following actions:

• Perform a traceroute to find out where the failure is occurring.
• Review the Routing Table.

If the host is unreachable, a possible result is Hop Count Exceeded, Network Unreachable, or Host Unreachable.

If the hop count is exceeded and the IP address column on the last entry or entries is an *, a device is not responding (possibly the destination host). If so, or if the network or host was unreachable, examine the Routing Table or Interfaces for the last listed hop. Apply I (Interfaces) or R (Routing Table) against the host or router that is recorded as the last hop on the list.

When using the Routing Table action, your TCP/IP management region uses masking. Masking ensures that the presented Routing Table contains only those addresses that are relevant to the path you are tracing. This mask is a network address mask that is based on the class of address being traced. For example, if you are tracing a route to the foreign host 192.168.2.66, your region inserts the mask 192.168.2.*.

If the hop count is exceeded and the same number of hops is shown in the Hop Limit field, increase the hop limit and reissue the traceroute by pressing F6 (Action).

If the ping indicates a slow response, you can review the following information in your traceroute results:

• Review the Notes column to see if there are any outstanding alerts for the node. If the IP Node Monitor is monitoring the node, look for any alerts are outstanding.
• Review the trip times—these times usually increase at a steady rate the further along the route the packets are sent. A sudden and marked increase in the trip time for a particular hop along the route indicates a possible problem at that hop. To investigate further:
  • Apply the I (Interfaces) action to the host or router at the hop where the problem is occurring, and investigate the Interfaces for clues to the performance degradation.
  • Apply the TN (Telnet) action to start a Telnet connection to the host or router where the problem is occurring, and investigate the configuration or other possible causes.
• Review the hop list—if there is no evidence of slow trip times, look at the hop list for an unexpected route.

Note: To analyze response time problems, use the Transaction Path Analyzer facility.

If the traceroute does not return any abnormal results, use the reporting facility to look for large file transfers.