[Previous Section] [Back to Table of Contents] [Next Section]

Personal System Administration Guide


Chapter 8
Troubleshooting

This chapter contains detailed troubleshooting information on these topics:


Responding to System Monitor Warnings

The System Monitor keeps track of system activity, and notifies you when a critical error has occurred or is about to occur. It also provides online help that often gives you enough information to solve the problem. Always respond to these messages as quickly as possible to avoid losing valuable information.

To change the way the System Monitor notifies you of problems, or to turn off notification altogether, use the System Error Settings control panel; to start it, click the words System Error Settings now.

To get more detail on a particular message, you can open the System Log Viewer by choosing "View System Log" from the System toolchest, or by clicking the words View System Log now.


Troubleshooting Disk Space Problems

For information on managing and freeing up disk space, see Chapter 6, "Managing Disk Space."


Troubleshooting Problems with Removable Media

When you insert a floppy disk, floptical disk, SyQuest disk, CD, or tape into a drive, the drive's icon should change to show that it recognizes the new media.

If the icon doesn't change within one minute of inserting the media, eject the media using the hardware eject button, and insert it again. (The system sometimes takes up to 30 seconds to recognize tape media.) If the system still does not recognize the media, save your work, and restart the system by choosing "Restart System" from the System toolchest. After you log in, the system should show the inserted media.

If the floppy icon changes to generic drive icon, or if you ever double-click the floppy icon and see an error message that says "Unknown Device", there is an unformatted floppy or floptical disk in the drive. To format it, eject the disk, then see "Formatting Floppy and Floptical Disks".

If you are trying to copy files onto a disk that's formatted for DOS files and you get an error message that says "I/O Error", your file names are too long or do not conform to the DOS naming conventions. DOS file names can contain no more than eight characters, a period (.), and a three character extension (for example, projects.exe uses the maximum file name length).


Rebuilding System Setup Information

In very rare cases, the data that the system stores to identify users, systems, peripherals, and system setup information may become corrupted. If this happens, you'll see an error message that reports that system setup information has been lost or damaged, and refers you to this section on rebuilding the information.

You should follow the steps below only when you see such an error message. After you rebuild the information, no users will have administrative privileges except for the Administrator, and the system will have no Primary User. You need to use the User Manager to designate a Primary User and each Privileged User.

To rebuild the information, follow these steps:

    Log in as root through a shell window.


    Type:

    /etc/init.d/mediad stop

    /etc/init.d/cadmin stop

    /etc/init.d/cadmin clean

    /etc/init.d/cadmin start

    /etc/init.d/mediad start

    Wait for about five minutes, then try the operation again.


Troubleshooting Shared Resources Problems

If you cannot share resources, or cannot drag shared resources from other systems on to your own desktop, either the optional NFS software is not installed or is not turned on, or an important NFS utility called automount probably is not turned on.

Only the Administrator can turn NFS and automount on since you need to log in as root. Follow these steps:

    Log in as root through a shell window.


    Check whether NFS is installed and turned on by typing:

    chkconfig | grep nfs


    Type:

    chkconfig automount on

    /etc/init.d/network stop

    /etc/init.d/network start

    Close the shell window by typing:

    logout

    Restart the system by choosing "Restart System" from the System toolchest.


Troubleshooting Network Errors

When you see error messages when you try to perform an operation that uses the network, wait for a few minutes, then try the operation again. The network may just be temporarily overloaded.

Troubleshooting General Network Errors

If the operation cannot succeed after a few tries, test the connection by following these steps:

    Choose "Unix Shell" from the Desktop toolchest.

    Use the /usr/etc/ping command with the hostname of a system that appears in your Networking tool. For example, if the remote hostname is mars, type:

    /usr/etc/ping mars

    Then press <Enter>.


    Check the console window for error messages.


    A Privileged User should make sure TCP/IP is turned on, and, if yours is an NIS network, make sure NIS is turned on and the correct NIS domain name is entered.

Troubleshooting Network Errors in the Desktop

If the network connection is working, but you still cannot perform a network operation (such as searching for remote drives), check to see whether the objectserver is running . Follow these steps:

    Choose "Unix Shell" from the Desktop toolchest.

    In the shell window, type:

    ps -ef | grep objectserver

    You should see three lines whose last columns show:

    /usr/Cadmin/bin/objectserver
    /usr/Cadmin/bin/objectserver
    grep objectserver


    If the objectserver is running on your system, either the objectserver is not running on the remote system that you're trying to access, or no directoryserver is running on your local network.


Troubleshooting Standard Printing Problems

Once you've used the Printer Manager to set up the printers that you want to access, printing files is usually very straightforward - you ask an application to print a file, then pick up your completed job from the printer.

The work that the system does to make files print, however, is fairly involved - especially when you are sending files over a network to printers that are not directly connected to your workstation. If printing problems do arise, the information in this section should help you correct them quickly.

The information applies to you only if you are printing from within an application or from the desktop.

If you are printing from the command line using lp, see "Using the lp Spooler" in the optional IRIX Advanced Site and Server Administration Guide for more troubleshooting information on lp. If you're using lpr, see the lpr man page for information on basic use.

This section covers two main topics:

A Troubleshooting Roadmap

A successful printing process includes these four basic steps:

See "Understanding the Printing Process" for a more detailed description of what the system does at each step.

The troubleshooting steps below show you how to determine which part of this process is failing. Be sure you know the Administrator's password; various steps require that you become the Administrator.

    If you sent your job to the printer more than 30 minutes ago, send it again. This way you can monitor its progress from the start.

    Start the Printer Manager by choosing "Printer Manager" from the System toolchest.

    Find the icon for the printer to which you sent your job.


    Find your print job in the queue.

    The entries in this window are jobs that have already reached the printer's queue.


    Physically go to the system to which the printer is connected. Its queue contains all print jobs that have actually reached the system. Find your print job in the remote queue.

    Your job is labeled with your login name and is the same size as it is in the local queue. It does not have the same job number.


    Watch the queue of the printing system (the system to which the printer is attached).


    Make sure the printer is printing requests. Physically go to the system to which the printer is attached, open the printer's queue window, and choose "Printer Printing Your Queued Jobs" from the Queue menu.


    Check whether lpsched, the print spooler that controls the flow of jobs out of the queue, is running. On the printing system, open a shell window and type:

    lpstat -r

    Then press <Enter>.


    Check the physical state of the printer; for example:


    Remove all jobs from the queue and choose "Send Test Page" from the Printer menu to send a test page.

Job Never Appears in the Local Queue

Use this section if the printer icon to which you sent your print job appears in the Printer Manager, but your job does not appear in the local queue. You should be looking at the printer's queue window.

    Choose "Send Test Page" from the Printer menu to test the printer setup.


    Make sure you followed the correct steps to specify a particular printer for your job.


    Make sure you gave the application all the printing information it needs.

    Some applications (such as IRIS ShowcaseTM) have both a print command and a print dialogue box. If a dialogue box appeared and you didn't notice it, didn't fill it out correctly, or didn't confirm your information (that is, click an Accept or OK button), the job will not go to the queue.


    Open the console window and check for error messages.

    Note: You must look in the console; error messages will not be reported to any other shell window.


    Look in the file /var/spool/lp/log for error messages.

Job Never Appears in the Remote Queue

Use this section if your print job appears in your local queue, but does not appear in the queue when you view it on the system that's connected to the printer. You should be looking at the printer's queue window.

    See if there is already a different job in the remote queue that was sent from your local queue (that is, a job that belongs to you or to another user on your local system).


    Make sure the local printer queue is sending print requests to the remote system.


    Make sure the information about the remote system and remote printer that is shown in the printer's queue window is accurate.


    Check whether lpsched, the print spooler that controls the flow of jobs from the local queue to the remote system, is running. On your own system, open a shell window and type:

    lpstat -r

    Then press <Enter>.


    Test the network connection by opening a shell window and using the /usr/etc/ping command with the remote system's hostname. For example, if the remote hostname is mars, type:

    /usr/etc/ping mars

    Then press <Enter>. You see some messages that will repeat indefinitely; to stop the messages, press <Ctrl-C>. You see a summary of the connection. Look for these lines:

    mars PING statistics
    <#>packets transmitted,<#> packets received,0% packet loss


    Check the access permissions on the remote system by trying to copy a file to the remote system using the same login account that lpsched uses to copy over your job. For example, use jot to create a small text file named testit, then copy it to the remote system (mars) using the lp account; type:

    su lp

    rcp testit lp@mars:/usr/tmp


    Choose "Send Test Page" from the Printer menu on your own system to test the printer setup.


    Physically go to the remote system and open the console window to check for error messages.

    Note: You must look in the console on the remote system; error messages will not be reported to any other shell window.

Job Disappears from a Queue but Never Prints

Use this section if your print job disappears from a local or remote queue but the printer never prints it out.

    Check your mail messages.


    Check whether the printer is printing any jobs.


    Send the job again, and check whether the printer receives it.

    Most printers have a status mechanism (a blinking light or digital message) that shows that the printer has received a job and is trying to print it.


    Remove all jobs from the queue on the printing system and choose "Send Test Page" from the Printer menu to send a test page.

Checking and Restarting lpsched

Use this section to check whether lpsched is running, and to restart it if necessary.

    Check whether lpsched is running by typing:

    lpstat -r

    Then press <Enter>.


    Turn on lpsched if it is not running.


    Make sure lpsched is now running by typing:

    lpstat -r

    Then press <Enter>.

Understanding the Printing Process

This sequence of steps describes the process that your system uses to print files. The details (such as system name, application name, and job IDs) are only examples.

    On your own system (saturn), you ask an application (IRIS Showcase) to print a file (slide1), and explicitly or implicitly request a particular printer (color-seiko).

    Showcase (or another filter program) creates a new version of slide1 (a new file) that is in the correct format for color-seiko.

    Showcase runs the lp command on the file on saturn. lp assigns the file a job ID number (10), sends it to color-seiko's queue, and alerts lpsched (the spooler that controls the flow of jobs out of the queue) that the file (job #10) is ready to be printed. The printer's queue displays job #10 in the local queue for color-seiko.

    color-seiko is actually connected to another system on the network (mars) where it is named seiko1. When job #10 reaches the top of color-seiko's queue, lpsched copies it across the network to the /usr/tmp directory on mars.

    lp on mars makes a copy of job #10 in /var/spool/lp/request. It then assigns it a new ID number (20) that doesn't conflict with other IDs on mars, sends it to seiko1's queue, and alerts lpsched on mars that job #20 is ready to print. The Printer Manager on mars shows job #20 in seiko1's local queue; the Printer Manager on saturn continues to show job #10 in color-seiko's queue.

    When job #20 reaches the top of the queue, lpsched sends the job over a cable to seiko1.

    seiko1 receives job #20 and prints out slide1 on paper or a transparency.

    Job #10 disappears from color-seiko's queue; job #20 disappears from seiko1's local queue. lpsched sends the next job in color-seiko's local queue to mars.

Where the Process May Fail

This section shows how the process may fail at each step shown in "Understanding the Printing Process"; it does not describe how to correct the failure. See "A Troubleshooting Roadmap" for a step-by-step approach for isolating and correcting failures in the printing process.

    When you ask an application to print a file on a certain printer:


    When the application or filter tries to create a new version of the file:


    When the application runs lp on the file:


    When lpsched tries to copy the file to the remote system:


    When lp on the remote system tries to process the file:


    When lpsched sends the job over the cable to the printer:


    When the printer receives the job and tries to print it:


    When jobs disappear from the queue:

Troubleshooting Tools

When you troubleshoot printing problems using "A Troubleshooting Roadmap," you use a number of different tools and techniques. This section summarizes the tools, and suggests other sources of information on isolating and correcting problems.

"Send Test Page"


This choice appears in the Printer menu in the Printer Manager window and in each individual printer's queue window. When you choose "Send Test Page," you run the lp command on a sample file that is already in the correct format for that printer; in other words, you bypass steps 1 and 2 of the printing process described in "Understanding the Printing Process".

lpstat

This IRIX command provides an alternate view of the queues that the Printer Manager displays; it also reports whether lpsched is running. To see a full lpstat listing that includes queues for every printer on the system plus the status of lpsched, you would type: lpstat -t

/var/spool/lp/log


This file contains a history of all printing activity and errors. Its messages are often difficult to understand, but you may find some useful error information that you can use for troubleshooting. For example, a "login incorrect" message that has a timestamp near a time when your job could not reach a remote queue may make you suspect that there is a password on the lp account on the remote system.

/etc/init.d/lp start


This IRIX command restarts the scheduler, lpsched.

Other documentation


Chapter 9, "Administering Printers" in the optional IRIX Advanced Site and Server Administration Guide contains a detailed table of printing error messages and describes how to use the lp system from an IRIX shell; the lp man page gives a complete listing of all the command line options for lp.


Troubleshooting lpr Printing

This section shows you how to diagnose and fix these common problems:

If the print request never reaches the printer's queue (if you don't see the request when you type lpq), follow these steps:

    Make sure you entered the print command correctly.


    Check for error messages.


    Send a simple file to the printer, such as /etc/group.

If the print request reaches the queue, but never disappears from the queue because it cannot reach the remote host (printing system), follow these steps:

    Check for error messages:


    Make sure you can access the printing system.


    Ask the Administrator of the printing system to verify that your system's hostname is in the /etc/hosts.equiv file on the printing system and that your system's hostname and IP address are in the /etc/hosts file. Also ask the Administrator to verify that the printer is enabled and ready to print requests. When this is done, try printing again.

If the print request reaches the printing system and disappears from the queue, but the printer either never prints it or prints something unexpected, follow these steps:

    Log in as root through a shell window.


    Send a test job to check whether lpr on your system is spooling jobs correctly.


    Once you know that lpr is working correctly on your system, you can assume that there is a problem with the printing system. Contact the Administrator of the printing system; this person will have to perform some of the steps in the rest of this troubleshooting procedure.


    Once you know lpr is working correctly on the printing system, the Administrator should restart the printer (named colorful) on that system by typing:

    /usr/etc/lpc start colorful

    Then press <Enter>.


Software Installation

For troubleshooting information, see "Troubleshooting Software Installation."


Running Confidence Tests

If any of your physical devices are not working properly, run the appropriate confidence test. Follow these steps:

    Choose "Run Confidence Tests" from the System toolchest.

    Double-click the device that you want to test. On-screen instructions guide you through the test.

    If a device is faulty, contact your local service organization.


[Previous Section] [Back to Table of Contents] [Next Section]


Send feedback to Technical Publications.

Copyright 1997, Silicon Graphics, Inc. All Rights Reserved. Trademark Information