Troubleshooting IPX/SPX
The following sections describe:
If you suspect that the IPX protocol is not operating correctly, you can use the following commands to verify its state:
For IPX/SPX to operate correctly, NPSD, SAPD, and NVTD must be running.
To make sure that the daemons are running,
execute the following command lines:
ps -ef | grep npsd
ps -ef | grep nvtd
ps -ef | grep sapd
If there is no output from these commands, these daemons are not running.
To restart the daemons,
execute the following command line:
/etc/ipx restart
If services on the network are inaccessible, then SAPD may not be operating correctly. Use showsvcs to display the services. If services are not displayed, then IPX/SPX may not have been configured correctly.
If LANs on the network are inaccessible, then routers on the network may not be operating correctly. To verify the operation of routers, use the drouter command. If known LANs do not appear in the output of drouter, it is likely that the routers for the missing segments are not operating. For further details, refer to the documentation that came with your routing hardware.
Incorrectly configured LAN adapters are the principal cause of IPX failures. It is essential to verify that all LAN adapters have been installed correctly without any hardware and/or software configuration conflicts. Examine settings for: IRQ, I/O port, DMA, and shared-memory settings (depending on the type of adapter). For further details, refer to Appendix A, ``Configuration parameters'' in Configuring Network Connections, as well as to the documentation that came with your network adapter hardware.
The llistat(ADM) command, provided with the network adapter drivers, can verify the operation of drivers. Execute the following command line:
llistatIn the llistat output, In/Out counts of zero typically point to configuration or installation errors.
The SCO IPX/SPX installation mechanism prevents many potential installation errors. However, it is important to have a clear picture of the network topology prior to software installation. This is particularly important when routers and multiple subnetworks are being configured.
Incorrect specification of internal or subnet network numbers cause host addressing problems. All server network numbers and subnet numbers have to be manually checked for uniqueness. This should be done prior to configuration or installation. It is advisable to establish a numbering convention to distinguish internal (that is, logical) networks from physical subnetworks. This facilitates the interpretation of the outputs from certain utilities, for example, drouter.
The naming of servers, as well as services offered by servers, should be done in a meaningful way. Duplicate names and addresses can cause serious addressing problems. A meaningful naming scheme and corresponding addressing convention reduce potential problems significantly.
Each NVT connection requires at least one pseudo-tty to service it. If you are experiencing a high number of failed connections, make sure that the number of configured pseudo-ttys is not less than the value of the nvt_max_logins parameter in NPSConfig. See NPSConfig(SFF) for more information.
If you use a hot-key to bring up DOS when using NVT.EXE for several minutes at a time, the UNIX session may be disconnected. This occurs because the NVT protocol specification states that new connection requests will use existing inactive connections.
Increase the value of the parameter NVT_KEEP_ALIVE_INTERVAL_TIME in nvt_tune.h to work around this problem. Once you have increased this parameter, you will need to relink the kernel and reboot the system. For more information, see the nvt_tune.h(SFF) manual page.
If the changes made during the reconfiguration
of IPX/SPX do not seem to take effect,
you may have omitted relinking the kernel and/or
restarting IPX/SPX.
When the IPX/SPX configuration is changed
using the Network Configuration Manager, you are informed
when it is necessary to relink the kernel and when
it is simply enough to restart IPX/SPX.
After a configuration change, make sure that the UNIX kernel has
been relinked and the system is rebooted or, if relinking is not required,
that IPX/SPX is restarted with the command line:
/etc/ipx restart