[Digital logo]
[HR]

DECnet-Plus
Problem Solving


Previous | Contents

2.3 Understanding Your Network Topology

There are many ways to collect network topology information. Two methods are:

2.3.1 The decnet_migrate Tool

You can use the decnet_migrate tool to get the following information for each node in your network:

You can also use the decnet_migrate show path function to trace network routes between one node and another.

2.3.2 When to Use the decnet_migrate Tool

Use this tool when you need a detailed map of parts or all of the complete network topology. If you have routers that use the Simple Network Management Protocol (SNMP), the decnet_migrate tool cannot collect information about them. Use the tools those routers provide to collect network topology information.

Collecting this information with the decnet_migrate tool can take a significant amount of time depending on the options you select.

2.3.3 References

The DECnet-Plus network management documentation explains how to use the decnet_migrate commands collect, report, and show path in detail. Sections 2.8.1 and 2.8.2 describe which NCL and NCP commands to use to trace network paths in the DECnet-Plus environment.

2.4 Recognizing Node Names and Synonyms

It is useful to have node name information when trying to isolate faults. Node names reflect either a Phase IV or DECnet-Plus style.

2.4.1 Identifying Node Names

Table 2-3 shows the differences between Phase IV and DECnet-Plus node names.

Table 2-3 Phase IV and DECnet-Plus Node Name Differences
Characteristics Example
Phase IV node names are:
  • Up to six alphanumeric characters
  • Stored in a local network database
MYNODE
DECnet-Plus node names are:
  • Typically, longer than six alphanumeric characters
  • Divided into the following elements:
    • Namespace nickname for DECdns or Local for the Local namespace
    • Directory names
    • Node name
  • Stored in the DECdns distributed namespace or the Local namespace.
ABC:.eng.node1

2.4.2 Finding Node Names and Node Synonyms

Use the following commands to find a system's node synonym or to find a DECnet-Plus full node name from a system's node synonym:
To Find A: Use This decnet_register Command:
Phase IV synonym decnet_register>show node node-id
Full node name decnet_register>show node node-id

2.5 Finding the Operating System and Version

Do the following to find your system's current operating system and version:

  1. Log in to the system and look at the system prompt.
    If the Prompt Is: Then the Operating System Is:
    Username: OpenVMS
    login: Digital UNIX
  2. If the operating system version does not appear when you log in, enter one of the following commands:
    If the Operating System Is: Enter:
    OpenVMS $ show system
    Digital UNIX # strings /vmunix|grep '(Rev')

2.6 Getting DECnet-Plus Circuit Information

To find basic routing circuit information, use Network Control Program (NCP) commands for Phase IV nodes and Network Control Language (NCL) commands for DECnet-Plus nodes. Refer to your NCP documentation for information about NCP commands.

2.6.1 Routing DECnet-Plus Circuit Information for Problem Solving

The following information can be useful when you need to solve routing circuit problems:

2.6.2 Finding Routing Circuit Names, Types, States, and Adjacencies

Do the following to find a circuit name, type, and state:
To Find Routing Circuit: Enter This NCL Command:
Name ncl>show node node-id routing circuit * name
Type ncl>show node node-id routing circuit * type
State ncl>show node node-id routing circuit * state

Routing circuit adjacencies result from nodes exchanging identification information. Finding circuit adjacencies is a quick way to:

Routing circuit adjacencies exist for wide area network (WAN) circuits, and local area network (LAN) circuits. Use the following NCL command to find routing circuit adjacencies:

ncl>show node node-id routing circuit circuit-id -
_ncl>adjacency * all status

2.6.3 Considerations for Broadcast Circuit Adjacencies

When you look at broadcast circuit adjacencies, the output can be extensive because many adjacencies can exist at one time. It can be more useful to request certain types of information rather than all information.

For example, you could use the following NCL command:

ncl>show node node-id1 routing circuit circuit-1 adjacency -
_ncl>node-id2 type

2.7 Preparing to Find a Network Path

Tracing a path from one end system to another can isolate network reachability problems and also can provide network topology information. The information in this section describes the type of information you need before you trace a network path.

2.7.1 Types of Network Information to Find

Use the Network Control Program (NCP) to get network path information from Phase IV nodes. Use the Network Control Language (NCL) to get network path information from DECnet-Plus nodes. Table 2-4 lists the information you need to trace a network path and how to identify this information.

Table 2-4 Types of Network Information to Trace
Information Description
Node Addresses Nodes have addresses that DECnet uses when sending data through a network. Phase IV nodes only have one node address; DECnet-Plus nodes can have up to six NSAP addresses; three that NSP use and three that OSI Transport use.

Phase IV node addresses and NSAP addresses are different. However, you can translate a Phase IV node address into an NSAP address.
Phase IV Node Addresses: NSAP Addresses:
Are called node numbers. Are called network service access points (NSAPs).
Are 2 bytes in length. Can be up to 20 bytes (40 hexadecimal digits) in length.
Contain an area number
(1 -- 63) and a node number (1 -- 1023).
DECnet-Plus Phase IV--compatible addresses contain:
  1. Phase IV prefix (the IDP--Initial Domain Part).
  2. Local area number - this is equivalent to a Phase IV area number.
  3. ID; a 6-byte field that is formatted like a 48-bit IEEE address and uniquely identifies a node within its area. This is equivalent to a Phase IV node number.
  4. A 1-byte NSAP selector that identifies the user of the Network layer, either NSP or OSI transport. The hexadecimal number %x21 indicates OSI transport; the hexadecimal number %x20 indicates NSP.

Network entity titles (NETs) Used to identify a node when it is not necessary to identify the transport software in use. A NET has the same format as an NSAP with a selector byte of %x00.
System type A system can be an end system, a level 1 router, or a level 2 router.

2.7.2 Finding Node Addresses for Node Names

A Phase IV node has one address; a DECnet-Plus node can have multiple NSAP addresses. You can determine a node's address as follows:
For This Node Type: Enter:
Phase IV The NCP command:

tell node-spec show exec

DECnet-Plus The NCL command:

show node node-id routing port * nsap address

If you cannot get the remote node's address in this manner, you need to log in to that system directly.

2.7.3 Determining Network Entity Titles for DECnet-Plus Nodes

A network entity title (NET) has the same format as a system's network service access point (NSAP), except the last two digits are set to 00. For example, if the NSAP is 49::00-0D:AA-00-04-00-7F-34:20, the NET is 49::00-0D:AA-00-04--00-7F-34:00.

2.7.4 Converting Phase IV Addresses to NSAPs

If you know a system's Phase IV address, you can do the following to convert it to an NSAP (see your DECnet introduction documentation for details about NSAPs):
Step Action
1 Ensure that the NSAP local area is in the Phase IV area in hexadecimal notation. For example, the Phase IV area 1 becomes the NSAP local area 00-01, and the Phase IV area 63 becomes the NSAP local area 00-3F.
2 Convert the Phase IV node ID to the NSAP node ID:
  1. Use the formula (area* 1024) + ID to convert the Phase IV area and ID to a single decimal value.
  2. Convert the value from step (a) to a four-digit hexadecimal number, and swap the first and last pairs of the hexadecimal digits.
  3. Use the number from step (b) as the last four digits of the NSAP's node ID field, and add the prefix aa-00-04-00.

Example of Phase IV Address Conversion

In this example, the network IDP (initial domain part) is 41:45436192:, the DSP (domain-specific part) is local-area:node-id:20, the Phase IV address is 43.258, and the node uses the NSP transport. You create the NSAP as follows:

IDP and selector   -> 41:45436192:local-area:node-id:20 
43 decimal         -> 2B     hexadecimal (local area) 
(43 * 1024) + 258) -> 44290  decimal 
44290 decimal      -> AD02   hexadecimal 
AD02  swapped      -> 02AD   hexadecimal (node ID) 
 
Resulting NSAP     -> 41:45436192:00-2b:aa-00-04-00-02-ad:20 
          

2.7.5 Converting NSAP Addresses to Phase IV Format

If you know a system's NSAP, you can do the following to convert it to a Phase IV address:
Step Action
1 Check that the local area is less than or equal to 63 decimal or 3F hexadecimal and the node ID field begins with aa-00-04-00. If both of these conditions do not exist, then the NSAP does not contain a Phase IV address and cannot be converted. If it does, go to the next step.
2 Extract the last four digits of the node ID field.
3 Swap the last two pairs of digits, and convert the value to decimal.
4 Calculate the Phase IV area and ID values:
  1. area= value/1024
  2. id= value - (area * 1024)

If the calculated area value is not equal to the area value obtained from the NSAP's local area field, the NSAP does not contain a valid Phase IV address.

Example of NSAP Conversion

In this example, the NSAP is 37:81076541234:00-19:aa-00-04-00-62-64:21. You calculate the Phase IV address as follows:

 
19 hexadecimal (from local area)  -> 25 decimal 
62-64 (from node-id)              -> 6462 hexadecimal 
6462 hexadecimal                  -> 25698 decimal 
25698/1024                        -> area of 25 
25698 - (25 * 1024)               -> node ID of 98 
 
Resulting Phase IV address        -> 25.98 
 

2.8 Tracing DECnet-Plus Network Paths

You can trace a network path for a pure DECnet-Plus environment and for a mixed environment that has DECnet-Plus and Phase IV nodes.

2.8.1 Tracing the Network Path of DECnet-Plus Nodes

You can trace the path from one node to another with the following command:

$ run decnet_migrate 
DECNET_MIGRATE>sho path from 
NAMES:.NETA.NODEA to NAMES:.NETB.NODEB 

Do the steps in the following table to trace a network path from one DECnet-Plus end system to another:
Step Action
1 Find the destination NSAPs.
2 Analyze the NSAPs to find the area addresses of the destination system.
3 Find a DECnet-Plus router that the source node uses. Do the following:
  1. Find the source node's routing circuits. Use the following NCL command:

    ncl>show routing circuit * all

  2. Select a circuit whose state is ON, and use the following NCL command to display the NET of the associated router:

    ncl>show routing circuit circuit-id -
    _ncl>adjacency * all

    The LAN address in the display may show the actual node name of the router. You can use this or the NET in this procedure. A LAN address is not displayed in all situations. For example, if tracing a path from an X.25 DA circuit, no LAN address is displayed.

A DECnet-Plus router may not exist. See the procedure in Section 2.8.2 if the next node in the path is a Phase IV router. If the next node in the path is not a Digital router, see that system's documentation.

4 If the router you find in the previous step is in the same area as your destination node, go to the next step. If the router is in a different area than the destination node, do the following:
  1. Search the router's destination area database for the destination area. Use the following NCL command (on OpenVMS systems, the router-id is the router's node name or the NSAP without punctuation and prefixed with net$ -- for example, net$490013AA00040089FF20; on Digital UNIX systems, the router-id can be the router's NET or the node name):

    ncl>show node router-id routing -
    _ncl> destination area destination-area all


    Note

    You can also use the decnet_register utility to display information about node names.

  2. Look at the routing circuit adjacency that the previous command displays. Use the following NCL command (On Digital UNIX systems, the router-id can be the router's NET or the node name):

    ncl>show node router-id routing -
    _ncl>circuit circuit-id adjacency -
    _ncl> adjacency-id all

    This command displays the NET for the next router in the path. If the next router is not in the same area as the destination node, repeat steps (a) and (b) until you locate a router in the same area.

5 When you find a router located in the same area as the destination node, do the following:
  1. Search the router's destination node database to find the circuit adjacency that the router uses to reach the next node in the path. Use the following NCL command:

    ncl>show node router-id routing destination -
    _ncl>node node-id all

  2. Show the routing adjacency to determine the next node in the path. Use the following NCL command:

    ncl>show node router-id routing circuit -
    _ncl> circuit-id adjacency adjacency-id all

  3. If the adjacency leads to another router, repeat steps (a) and (b) until you reach the adjacency that connects to the destination node.

2.8.2 Tracing a Network Path in a Mixed Environment

Your network can contain DECnet-Plus and Phase IV nodes. If you start to trace a path from a DECnet-Plus node and the next router in the path is a Phase IV router, do the following:
Step Action
1 Exit NCL and invoke NCP (or invoke NCP at the NCL prompt).
2 Use NCP to get the Phase IV node address of your destination node and the router.
3 Find the next node in the path. Use the following NCP commands (the router-id can be the node name or the Phase IV address):

ncp>tell phase_iv-node-id show node -
_ncp> destination-node-address

4 If the next node in the path is not the destination node, repeat the commands in step 3, using the next node that the show node command displays as the next-node-id, until you reach the destination node.
5 If NCP returns an error message, the next node in the path could be a DECnet-Plus node. Do the following:
  1. Exit NCP and invoke NCL.
  2. Repeat steps 4 or 5 in Section 2.8.1.

2.8.3 X.25 DA Circuit Considerations

If you are tracing a network path from an X.25 DA circuit, you cannot use NSAP or NET information to find the next node in the path. Use the following procedure to help you trace a path from an X.25 DA circuit:
Step Action
1 Find the reachable address subentity that has a prefix that best matches your required destination NSAP.
2 Select a DTE from the DTE Addresses attribute and determine the node address (NSAP or Phase IV synonym) associated with the remote DTE address.
3 If you cannot determine the node address, try to log in to the remote node using X.29 (PAD) and continue tracing the network path from the remote node.

2.8.4 NCL Takes Long Time While Translating Addresses to Names

When NCL displays a node address in response to a SET or SHOW command, it uses the services of DECdns to translate the address into a node name and displays the name along with the address.

If there is a problem with accessing a remote name server that prevents the translation from completing, NCL may take a long time translating addresses to names. If this happens, enter a Ctrl/Y to terminate NCL, then make the following definition prior to executing NCL:

$ DEFINE NCL$ENVIRONMENT NOBACKTRANS 

This causes NCL to bypass the address-to-name translation. To use this option on a systemwide basis, add this logical definition (with the /SYSTEM qualifier) to SYS$MANAGER:NET$LOGICALS.COM.


Chapter 3
Testing Network Reachability

This chapter describes the network reachability tests you can use in the DECnet-Plus environment.

Topics In This Chapter

The topics in this chapter are:

3.1 Types of Network Reachability Tests

You can use the following network reachability tests:
Test Description
Quick reachability Provides a fast indication that a remote node is reachable using applications such as dlogin or set host with the node address instead of the node name. If you cannot connect to a node using its node name but can connect to it using the address, the cause could be a naming service problem.

For Digital UNIX systems, you can also use OSI Echo function (OSI ping).

Loopback Loopback tests (node level, circuit level, and LAN with LLC test messages) let you thoroughly use network software and hardware by sending data through various network components and returning that data to its source for comparison.
DECnet Test Sender and Receiver ( dts/dtr) Throughput tests that allow you to load test the ability of different systems to exchange data. The dts/dtr tests are also useful if you want to check Transport layer connections.

Parameters are available to regulate such variables as message length, test duration, and type of data used.

X.25 and OSI Transport IVP (OpenVMS systems only) These tests check if the X.25 or OSI transport software is working correctly. Your installation documentation describes them.

3.1.1 Types of Loopback Tests

You can use the loopback tests described in this section for DECnet-Plus systems. Refer to your DECnet Phase IV documentation if you need details about similar tests for Phase IV systems.
Test Type Description
Node level Checks the logical link capabilities of a node by exchanging test data between DECnet tasks in two different nodes or in the same node. You use NCL to run this test, which enables you to connect to a loopback mirror application. The types of node-level tests are:
  • Local-to-Local loopback tests, which evaluate the local node's DECnet software using an internal logical link path; no physical device is used.
  • Local-to-Remote loopback tests, which verify network operation between the local node and a remote node.
Circuit level Checks a DECnet circuit by looping test data between a mop task and the communication hardware on one or two systems.


Previous | Next | Contents | [Home] | [Comments] | [Ordering info] | [Help]

[HR]

  PS_PROFILE_001.HTML
  OSSG Documentation
   2-DEC-1996 12:34:12.26

Copyright © Digital Equipment Corporation 1996. All Rights Reserved.

Legal