C H A P T E R 7

Troubleshooting

This chapter includes information about troubleshooting Sun Remote System Control (RSC) problems, and troubleshooting the server using RSC.

Troubleshooting RSC Problems

Can't log in to RSC

Check the RSC device name you are connecting to (Sun recommends the name servername-rsc).

Check that you are using your correct RSC username; it may not be the same as your system username.

Check that you are using your correct RSC password.

Can't connect to RSC using telnet

RSC supports a total of four concurrent telnet sessions per server, including command-line interface sessions and a graphical user interface (GUI) connection to the server console (Open Console). In addition, RSC supports up to four active concurrent RSC GUI sessions.

When the maximum number of telnet sessions are active, further attempts to connect using the telnet command will receive a connection closed error. The following example shows system messages for the UNIX operating environment:

% telnet bert-rsc

Trying 129.148.49.120...

Connected to bert-rsc.

Escape character is '^]'.

Connection closed by foreign host.

Note - Sun Enterprise 250 servers support two concurrent telnet sessions and three active concurrent RSC GUI sessions.

Can't connect to RSC through the Ethernet connection

First, log in to the server as root and see if the command rscadm status succeeds. If it does, RSC is working and there is an Ethernet configuration problem. Use the rscadm show command to check that Ethernet configuration variables are set correctly.

You can also:

Log in to RSC through the RSC serial port or modem and use the GUI or the shownetwork command to see the current settings.

Log in to another machine on the network and use the ping command to see if RSC is operating. Use RSC's name (for instance, servername-rsc), not the server name, as the argument to the ping command.

Run SunVTS diagnostics to check the Ethernet connection. The external Ethernet test requires that the device be connected to a functional 10-Mbit hub.

Run SunVTS diagnostics to check the RSC card.

Use the rscadm subcommand rscadm status to check RSC status.

Can't connect to RSC through the modem

First, log in to the server as root and see if the command rscadm statussucceeds. If it does, then RSC is working. Use the rscadm show command to check that modem configuration variables are set correctly (for Sun Enterprise 250 servers, check that serial port variables are set correctly).

As an alternative, you can also log in to RSC through the serial port and use the show command to check that the modem configuration variables are set correctly.

If the problem is with the modem:

Log in to the server as root and use the command rscadm modem_setup
to issue attention (AT) commands to see if you can place a call to an internal or external phone. This exercise will help to verify that the modem is working and that there are no problems with phone lines.

Run SunVTS diagnostics to check the modem.

Run SunVTS diagnostics to check the RSC card.

Note - For Sun Enterprise 250 servers, it is very important that the external modem is configured to use a fixed baud rate for incoming calls.

No alerts received from RSC

RSC does not send alerts for all RSC events. Check that the alert you are expecting is an event for which RSC sends alerts.

If email alerts are not received:

Check email configuration variables.

Check that the mail server and backup mail server are up and configured properly.

Try sending an email message to the recipients without using RSC.

Check Ethernet configuration.

If pager alerts are not received:

Check pager configuration variables.

Try sending a pager message by some method other than RSC.

Check with your pager service for proper pager configuration, or problems with the paging service.

Check that the RSC modem is operational. See Can't connect to RSC through the modem.

Alert messages may be too long for the pager or paging service used. Turn off verbose messages. See page_verbose.

RSC passwords are unknown

If users have forgotten RSC passwords or passwords are not working, log in to the server as root and use the command rscadm userpassword to assign new passwords. Inform RSC users of the new passwords.

Times in RSC event log do not match times in server console logs

RSC time is synchronized to the server's time each time the server reboots. To synchronize RSC time with server time at any time, reset the server, run the script /usr/platform/sun4u/sbin/rsc-initscript, or use the rscadm command rscadm date -s.

You can perform some RSC functions, but not others

Specific user permissions are required to perform functions. Check your permission level. In addition, the following problems may exist:

Cannot see console logs or access the server console using RSC.

Do one of the following:

Use the OpenBoot PROM diag-consolecommand to direct the console to RSC. (For Sun Enterprise 250 servers, use the diag-output-tocommand.)

Use the RSC command bootmode -u.

Choose Set Boot Mode using the graphical user interface and check the box labeled "Force the host to direct the console to RSC."

Cannot put the server in debug mode or use the RSC break command.

The server keyswitch is in the Lock position.

The poweroff command has no effect.

The server is already powered off.

The poweron command has no effect.

The server is already powered on, or the keyswitch is in the Standby position.

Troubleshooting Server Problems Using RSC

RSC is useful for troubleshooting a server that is not responding. If the server is responsive, connect to it as usual and use standard tools such as Sun Management Center, SunVTS, and OpenBoot Diagnostics.

If the server is not responding, log in to your RSC account and:

Check the RSC event log and server environmental status for problems.

Check console logs for recent error messages.

Try connecting to the server console to reboot the system.

Scenario: SIMM in memory bank 2 is causing continuous reboots

1. Log in to RSC and check the RSC event log.

You will see repeated server resets in the log.

2. Examine console logs.

You will see many SIMM bank 2 errors in the original run log, and eventually a panic message.

3. Connect to the server console.

You may see that the system is up now, but receiving periodic SIMM errors.

4. Log in to the server as root, and halt the system.

5. Enter the following OpenBoot PROM command to disable SIMM bank 2:

ok> asr-disable bank2

6. Reboot the server.

7. Order new memory for the server.