|
DECnet-Plus for OpenVMS Network Management
10.2 Manually Configuring MOP
The following sections show the commands to create the mop
entities you need to accomplish the tasks described in this chapter.
For the variables, substitute values appropriate to your configuration.
Refer to the DECnet-Plus Network Control Language Reference for more information about these attributes.
Section 10.2.2 shows how to set up the mop client for a network
server , while Section 10.2.3 shows how to set up the mop
client for an OpenVMS Cluster satellite.
10.2.1 Configuring MOP and MOP Circuits
This section shows how to create mop and mop circuits
entities. Figure 10-1 shows the mop entity and its
subentities.
Figure 10-1 mop Entity
- Create and enable the mop entity as follows:
ncl> create mop
ncl> enable mop
|
- Create a mop circuit entity and use the set
command to customize the entity definition for the specified circuit:
ncl> create mop circuit csmacd-0 type csma-cd (1)
ncl> set mop circuit csmacd-0 -
_ncl> link name csma-cd station csmacd-0, (2)
_ncl> known clients only value (3)
ncl> enable mop circuit csmacd-0 -
_ncl> functions (load server, dump server, console requester, -
_ncl> configuration monitor, loop requester, query requester, -
_ncl> test requester) (4)
|
- Specify any name you want for
circuit, but for convenience you might want to specify the
same name you used for the station entity (see next callout).
- For the data link station name, specify the
name of the circuit you used when you created the csma-cd
station entity. (For more information about the csma-cd
station, see Section 8.3.1.1.)
- Specifies whether MOP should attempt to
service load requests from remote systems that do not have a
corresponding client entity. Some network servers are designed
to request specific software by name. In such a case, there is no need
for a client entity to exist.
By default (false),
MOP tries to process requests for named software from unknown clients.
Set this attribute to true if you want MOP to ignore such
requests.
- Specify one or more functions for which you
want to use MOP. For instance, specify console requester if
you want to use the boot mop or load mop commands or
the console carrier.
Do not enable the configuration monitor unless
you plan to use it, because it consumes relatively large amounts of
system resources.
Note that for each mop circuit entity, DECnet-Plus
dynamically creates a mop circuit operation entity for each
ongoing MOP operation. Thus, you can issue the show command
for mop circuit entities to display current values of their
characteristics. For example:
ncl> show mop circuit circuit-name operation * all
|
10.2.2 Setting Up a MOP Client for a Network Server
You can have a variety of network servers on a network. Different kinds
of network servers require different levels of information for downline
loading and upline dumping. For instance, a terminal server requires
the hardware address and the system image. A DECnet Phase V router
requires the hardware address, the system image, and the management
image or script file, depending on the kind of router you have.
For information about configuring a mop client for an OpenVMS
Cluster satellite, see Section 10.2.3.
The following command example shows how to set up and define the
characteristics for the mop client subentity for a network
server. The mop client is a set of characteristics that
represents the target node. These collected characteristics are known
as the client database. The example shows all possible information you
might need to provide. Refer to your network server documentation for
the exact information you need to provide.
ncl> create mop client client-name
ncl> set mop client client-name -
_ncl> circuit circuit-name, - (1)
_ncl> addresses {lan-address, hardware-address}, - (2)
_ncl> secondary loader {filename}, -
_ncl> tertiary loader {filename}, -
_ncl> system image {filename}, -
_ncl> diagnostic image {filename}, -
_ncl> script file {filename}, -
_ncl> management image {filename}, - (3)
_ncl> verification octet-string, - (4)
_ncl> device types {device-types}, - (5)
_ncl> phase iv client address phase-iv-address, -
_ncl> phase iv client name phase-iv-name, -
_ncl> phase iv host address phase-iv-address, -
_ncl> phase iv host name phase-iv-name, - (6)
_ncl> dump file {filename} - (7)
_ncl> dump address address (8)
|
- Specifies the name of the MOP circuit over
which this client can be reached.
- Phase IV nodes can use an extended DECnet LAN
address as well as their hardware address, so you must include both of
these addresses in the addresses set.
Note
For clients configured with NET$CONFIGURE, the extended Phase
IV DECnet address is automatically added if you supply the Phase IV
client address when prompted. For clients configured with NCL, you must
manually include the additional address.
|
To calculate the extended Phase IV, proceed as follows:
- Convert the Phase IV node address (in the format
area-number.node-number) to its decimal equivalent,
using the following conversion algorithm:
(area-number
* 1024) + node-number
- Convert the decimal node address to its hexadecimal equivalent,
reversing the order of the bytes to form the hexadecimal node address.
- Incorporate the hexadecimal node address in the following format:
aa-00-04-00-hexnodeaddress
For example, to determine the Ethernet physical address of a node whose
node address is 63.171, calculate the following: (63 * 1024) +
171 = 64683 decimal = FCAB hexadecimal This calculation
results in the following Ethernet physical address of the node:
- secondary loader, tertiary loader, system
image, diagnostic image, script file, and management
image all specify files that can be used during a downline load.
The load sequence for a client-initiated request is as follows: The
first program to run at the target node is the primary loader.
Typically, this program is either executed directly from the client's
bootstrap ROM, or it is in the microcode of the load device. Usually,
the primary loader requests a secondary loader program, which then
might request a tertiary loader, which, in turn, might request an
operating system. The final module to be loaded is the management file
or the CMIP script file. In this sequence, each program or file
requests the next one until the management file or CMIP script file is
loaded. Refer to your network server documentation to find out
which of these files you need. In some cases, the client might specify
a file name when it requests to be loaded; in this instance, you do not
have to specify the file when setting up the client. Fields in
these file specifications are defaulted. If the file specification
comes from the client database, the default device and directory are
provided by the logical name MOP$LOAD. If the client specifies
a file name in the load request message, the default device and
directory are provided by the logical name MOP$NAMED_LOAD.
These logical names are created by NET$STARTUP,
but you can modify them. The default file type is .SYS for the
secondary loader, tertiary loader, system image, and diagnostic image
files. It is .DAT for management images and CMIP script files.
- Specifies the verification string. The value
is an octet string of up to 16 hexadecimal digits. Enter the value as
"%X" followed by an even number of digits. For more information about
specifying a verification string, see Section 10.2.2.1. The default is
%X0000000000000000.
- Specifies one or more device types
associated with this client. Use device types and omit
addresses if you want to set up a generic client
entity. The entity will be used for any incoming load or dump requests
that specify a matching communications device type. To discover the
communications device type for a particular network server, refer to
your network server documentation, or use the configuration monitor
function of MOP.
- Use phase iv client address, phase iv
client name, phase iv host address, and phase iv host
name to specify client and host names and addresses for Phase IV
systems. Enter these in the area-number.node-number
format.
- dump file specifies the file to
which MOP writes when it dumps the network server. Refer to your
network server documentation for information about dump support for
your server.
The default device and directory are provided by the logical name
MOP$DUMP. This logical name is created by NET$STARTUP
but you can modify it. The default file type is .dmp.
- dump address specifies the first
location of client memory to be dumped. Specifying any value other than
zero might affect your ability to use a dump analysis tool on the dump
file. Change the dump address only if your client documentation
suggests such action.
10.2.2.1 Setting Up MOP Service Passwords on a Network Server
To guard against unauthorized access, some network servers require that
a 16-digit hexadecimal service password be present in MOP load, boot,
or remote console requests. The way the password is set up for a
particular server depends on the design of that server. For example, a
server might use a console command to change a password held in
nonvolatile RAM.
When you attempt to boot the server using the NCL boot or
load commands, or to connect to the server's remote console
using the ccr command, you must specify the correct password,
or the server ignores the request. You can specify the password in the
verification client attribute or command parameter.
When using NCP in a Phase IV network, the service password is
interpreted as an unsigned integer of up to 16 hexadecimal digits. You
can omit leading zeros. In keeping with the definition as an integer,
the rightmost digits occupy lower addresses in memory than did the
leftmost digits.
When using NCL in a DECnet Phase V network, the service password or
verification value is interpreted as an octet string of up to 16
hexadecimal digits. You can omit trailing (rightmost) zeros on input.
Leftmost digits occupy lower addresses in memory than do
rightmost digits.
If a particular network server provides a command to set up the
password value, the password syntax might be like NCP, treating the
value as an integer, or it might be like NCL, treating the value as an
octet-string. If the server's syntax is like NCP, you need to specify
the password in different ways on the server and on the host system.
To convert an NCP-style value to an NCL-style value, do the following:
- Add leading zeros to make the value exactly 16 digits.
- Working from right to left on the NCP-style value, copy each pair
of digits to the NCL-style value.
- Prefix the NCL-style value with "%X".
For example, if you have an NCP service password value of:
The NCL verification value is:
Note
When using the Common Trace Facility (CTF) to trace MOP activity, the
trace shows the bytes of the verification value in the order in which
they are transmitted, which corresponds to the order in memory. For the
NCL verification value:
you would receive the following trace analysis output:
Verif=CD-AB-89-67-45-23-01-00
|
|
10.2.3 Setting Up a MOP Client for an OpenVMS Cluster Satellite
The following command example shows how to set up and define the
characteristics for the mop client subentity for an OpenVMS
Cluster satellite. The mop client represents the target node.
You can either use the commands listed below to set up your OpenVMS
Cluster satellite or you can use the command procedure
CLUSTER_CONFIG.COM. For more information about
CLUSTER_CONFIG.COM, refer to the OpenVMS Cluster Systems
for OpenVMS guide.
ncl> create mop client client-name
ncl> set mop client client-name -
_ncl> circuit circuit-name, -
_ncl> addresses {lan-address}, -
_ncl> system image {"@net$niscs_laa(dev:[root.])"}, - (1)
_ncl> verification octet-string (2)
|
- Specifies the load assist agent. The load
assist agent is an image that supplies MOP with the initial OpenVMS
load image for the OpenVMS Cluster satellite. DEV:[ROOT.]
indicates that you should specify the system root for the OpenVMS
Cluster satellite you are loading. For more information about OpenVMS
Cluster satellites, refer to OpenVMS Cluster Systems for
OpenVMS.
- Specifies the verification string. The value
is an octet string of up to 16 hexadecimal digits. Enter the value as
"%X" followed by an even number of digits. For more information about
specifying a verification string, see Section 10.2.2.1. The default is
%X0000000000000000.
Note
OpenVMS Cluster satellites do not use upline dumping; rather, they
write their memory image to the boot server's disk.
|
Section F.6 provides an example of configuring MOP for either an
OpenVMS Cluster satellite or a network server.
10.2.4 After Configuring MOP
After you have created the entities, defined their characteristics, and
enabled the mop client entities and the mop circuit
entities, use the load, boot, loop, query, and test
commands to perform MOP operations. (For information about using the
load and boot commands, see Section 10.5.1 and
Section 10.5.2.) For information about using the loop, test,
and query commands, refer to the DECnet-Plus Problem Solving guide. Use the
show command to display information about all aspects of the
mop entity and MOP operations.
10.2.5 MOP's Use of Default Directories
MOP uses logical names to identify directories where it expects to find
images or to write dump files. MOP defines the logical names shown in
Table 10-2 in NET$STARTUP.COM.
Table 10-2 MOP Logical Names and Default Definitions
| Logical Name |
Use |
Default Definition |
|
MOP$LOAD
|
Directory that MOP searches for files
specified for a client in the MOP client database, if they are
specified without a directory
|
SYS$SYSROOT:<MOM$SYSTEM>,
SYS$SYSTEM:
|
|
MOP$NAMED_LOAD
|
Directory that MOP searches for files
named in the
software id field of a MOP request program message issued by a
network server requesting a downline load
|
SYS$SYSROOT:<MOM$SYSTEM>
|
|
MOP$DUMP
|
Directory to which MOP writes a dump file in
response to a MOP request dump message issued by a network server
|
SYS$SYSROOT:<MOM$SYSTEM>
|
Phase IV DECnet used a set of MOM$XXX logical names.
Network server software already installed on your system and network
server installation kits that have not yet been updated for DECnet
Phase V expect these MOM$XXX logical names. For
backward compatibility, NET$STARTUP.COM defines these
MOM$XXX logical names to their default Phase IV
values.
Some network servers, however, create a separate directory for their
files. Network server instructions for a Phase IV DECnet environment
tell you to redefine a MOM$XXX logical name to
include that separate directory. For example, the DECserver 700 product
places its files in SYS$SYSROOT:[DECSERVER], and expects
MOM$LOAD to be redefined to
SYS$SYSTEM,SYS$SYSROOT:[DECSERVER].
For DECnet-Plus to load such network servers, you must redefine MOP
logical names to include the directories in which their files reside.
Network servers that do not have a MOP client database definition
always supply a software ed field in the request program
message they transmit when requesting a downline load. For these
network servers, redefine the logical name MOP$NAMED_LOAD to
include the directory. For example:
$ define mop$named_load sys$sysroot:[mom$system],sys$sysroot:[decserver]
|
For network servers that have a MOP client database definition, either
you can modify the MOP$LOAD logical name, or you can modify
the MOP client file specifications to include the directory
specification. For example, if the client files reside in
SYS$SYSROOT:[FOODIR], change:
ncl> set mop client fred system image {foobar.sys}
|
to
ncl> set mop client fred system image {sys$sysroot:[foodir]foobar.sys}
|
You must redefine these logical names after the network is started. If
you start your network during system boot, edit
SYS$MANAGER:NET$LOGICALS.COM to include these definitions.
DIGITAL recommends that you do not edit NET$STARTUP.COM, since
that file will be replaced by a new version in future releases of
DECnet-Plus.
10.3 Starting MOP
On an OpenVMS system, during the system boot
operation NET$STARTUP.COM executes. By default,
NET$STARTUP.COM does not start MOP. If you want MOP to start
automatically, you can define the logical name NET$STARTUP_MOP
by adding it to SYS$MANAGER:NET$LOGICALS.COM. For example:
$ define net$startup_mop true
|
If you do not start MOP automatically, you can start MOP manually any
time after starting the network by entering the following command:
$ @sys$system:startup network mop
|
This command creates the process portion of MOP, NET$MOP, in a
detached process. The command also executes the
NET$MOP_CLIENT_STARTUP.NCL and
NET$MOP_CIRCUIT_STARTUP.NCL scripts.
By default, MOP writes status messages to the file
SYS$MANAGER:NET$MOP_OUTPUT.LOG. You can override the default
log file by defining a logical name for it before starting MOP. The
following command defines NET$MOP_OUTPUT to the file name
SYS$MANAGER:MOP_APR95_STATUS.LOG:
$ define net$mop_output sys$manager:mop_apr95_status.log
|
|