2.1 SunDiag Main Window

When you start the SunDiag system exerciser, an OPEN LOOK window is displayed, as shown in Figure 2-2. The window is divided into four smaller sections:

• A system status panel that shows testing status.
• A performance monitor panel that displays performance statistics for the system under test.
• A control panel with buttons, settings, and window buttons for changing test parameters and options.
• A console window that displays operating system messages and test message. As superuser, you can execute operating system commands from this window.

The Options menu buttons are also noted in Figure 2-2; they are explained in the "Options Menu Buttons" section later in this chapter.

2.1.1 System Status Panel

The System Status panel is located in the upper left of the SunDiag main window, as shown in Figure 2-3. This window shows you know how the tests are progressing. The top of the system status panel shows:

• Current system status (idle, testing, or stopped)
• Number of successful system passes
• Total number of errors
• Elapsed time that testing has taken
• Page number in the current view (for example, 1 of 4)

You can reset all the values in these fields to zero by clicking the Reset button on the control panel.

The lower portion of the panel shows the number of successful passes and errors for each of the tests being run. The tests are grouped by device type, and an active test is marked with an asterisk. Use the Status View button to manipulate this list. (See Section 2.2.4.4, "Status View Button.")

The results displayed in this panel are stored in special log files. See the "Log Files Window Button" section in this chapter for details.

The "System passes:" field always shows the least number of passes that the longest test has completed. For example, if the pmem test has completed two passes while all other tests have completed four passes, the system status panel shows two system passes.

Figure 2-3

2.1.2 Console Window

The console window is displayed in the lower left corner of the SunDiag main window. This console window is a standard Solaris console window; you can enter commands from this window as superuser and display the results.

Using the -c option when starting the SunDiag program redirects system messages from your system console window to the SunDiag console window.

---

Note - Do not start another Solaris console window after starting the SunDiag program. If you do, the SunDiag window will stop receiving system console messages. The SunDiag console will continue to receive SunDiag messages, but not system console messages. Even if the new console window is closed later, the SunDiag console window will not receive console messages.

---

Figure 2-4

2.1.3 Performance Monitor Panel

The performance monitor panel is the same as that provided by the OpenWindows Performance Meter Deskset program; it provides a graphic display of system performance statistics.

Figure 2-5

The following table describes what each graph on the performance monitor panel represents:

Graph	Description
cpu	Percent of CPU being utilized
pkts	Ethernet packets per second
page	Paging activity in pages per second
swap	Jobs swapped per second
intr	Number of device interrupts per second
disk	Disk traffic in transfers per second
cntxt	Number of context switches per second
load	Average number of runnable processes over the last minute
colls	Collisions per second detected on the Ethernet
errs	Errors per second on receiving packets

2.1.4 Control Panel

The contents of the lower part of the control panel are determined by your system configuration. The SunDiag software automatically probes for available devices and displays them in this part of the control panel. Figure 2-6 on page 2-9 shows the control panel for a generic SPARCstation^(TM) system that contains:

• Physical and virtual memory
• Two SCSI disk drives
• Floppy disk drive
• Floating point accelerator
• Ethernet connection
• Serial ports
• Audio port
• Color graphics accelerator
• Special parallel port

When this window is displayed, check the devices listed in it against the devices you know are installed on your system. If there are any discrepancies, check the boot-up information in the /var/adm/messages file to make sure that the operating environment sees all of your hardware devices.

If the operating system does not recognize a hardware device that you know is attached to your system, check the hardware and the Solaris kernel configuration.

---

Note - When adding a new device driver in the Solaris operating environment, you must reboot the machine with the boot -r command before the SunDiag software will recognize the new driver.

---

When the software configuration agrees with the hardware configuration, you can begin to select settings for the devices you want to test.

In Figure 2-6, the tests shown with bold frames are enabled for testing.

2.1.4.1 Selecting Options and Clicking Buttons

Move the pointer to a button and click SELECT. Click buttons to perform actions such as Start/Stop and Reset/Suspend/Resume. Click window buttons (such as the Set Options and Log Files window buttons) to display a pop-up menu that contains additional controls.

Clicking SELECT on a setting chooses that setting. Selected settings are displayed with bold frames on monochrome monitors. On color monitors, selected settings appear shaded, as if they are "pushed in."

2.1.4.2 Dimmed Buttons

Some of the window buttons appear to be dimmed once testing has started. These buttons are not available until testing has stopped.

Figure 2-6

2.2 Setting SunDiag Options

The following sections describe the various buttons and settings that appear in the top part of the control panel.

2.2.1 Start/Stop Button

When you click the Start button, the SunDiag exerciser starts all enabled tests. Once testing has started, the wording on the Start button changes to Stop. Clicking Stop will stop testing.

---

Note - Click Stop only once. This button does not change back to Start until all tests have stopped, and some tests do not stop immediately.

---

The System status item on the system status panel changes from "idle" to "testing" once testing has started. When you stop testing, the system status changes to "stopping," then "idle."

2.2.2 Reset/Suspend/Resume Button

The Reset button resets the passes and error count for each test in the system status panel to zero. It also resets system passes, total errors, and elapsed time to zero. This button works only when testing has stopped.

Once tests have started, this button becomes Suspend. If you click Suspend, System status in the system status panel changes to "Suspended," and the Suspend button changes to Resume. When you are ready to resume testing, click the Resume button. After testing has stopped, the button changes to Reset.

2.2.3 Print Button

Click the Print button to take a snapshot of the current screen and print the file to the printer specified in the Set Options menu.

2.2.4 Log Files Window Button

The SunDiag system exerciser saves the status of its progress in three log files. These files contain Error, Information, and UNIX Messages. You can access these files by clicking the Log Files window button.

---

Warning - Do not attempt to edit message files while the SunDiag exerciser is running. You may only display, print, or remove these files. Editing these files while SunDiag exerciser is running causes the software to crash.

---

With the resulting pop-up menu you can to display, remove, or print each of those log files.

Figure 2-7

The three log files in this menu are:

/var/adm/sundiaglog/sundiag.err               SunDiag's Error Status Log
/var/adm/sundiaglog/sundiag.info              SunDiag's Information Log
/tmp/unix.msg                                 Solaris System Message Log

The sundiag.err file contains SunDiag test error messages and start/stop times. The status log file sundiag.info contains informative SunDiag and probe messages generated while starting and stopping the SunDiag program. The /tmp/unix.msg file is a concatenation of general UNIX messages in the /var/adm/message* files.

---

Note - Starting with the Solaris 2.1 operating environment, boot messages, SCSI error messages and some other debugging messages are no longer sent to the /var/adm/messages file. To continue to see these types of messages, make sure to boot your system using the boot -v (verbose) option.

---

Displaying Logs You can display any of the three log files by selecting the log file and then clicking the display button.

Figure 2-8

A pop-up test window will display the selected file (see Figure 2-9). Note that the log file name is printed at the top of the window.

Figure 2-9

Log File Test Message Syntax

TTT.VV.SSS.EEEE date time devicename testname [ERROR|FATAL|INFO|WARNING]
 message

Table 2-1

Arguments	Description
TTT	The test identifier.
VV	The two digit test version identifier.
SSS	The subtest identifier.
EEEE	The error identifier; the first digit is the priority identifier (message type) and the last three digits are the message identifier. The priority identifier number is one of the following: 9: ERROR 7: INFO 5: WARNING 3: FATAL
date time	Tells when the error occurred.
devicename	The device under test where the error occurred.
testname	The name of the test reporting the error.
message	This argument varies considerably, depending on the type of test running when the error occurred.

A typical error message generated by a test looks like this:

009.33.999.9025 12/13/89 18:54:49 xy0 fstest ERROR: not enough trace blocks on xy0

Refer to Table 5-2 on page 5-3 for a list of tests and their corresponding identification numbers.

The format of SunDiag kernel messages is:

TTT.VV.SSS.EEEE date time hostname sundiag [ERROR|FATAL|INFO|WARNING]: message

When the SunDiag kernel probes for devices at power-up, a message is displayed if any problems are found. The format is:

TTT.VV.SSS.EEEE date time hostname probe [ERROR|FATAL|INFO|WARNING]

Removing Log Files Click Remove on the Log Files window to truncate the selected log file (remove all but the most recent few entries of the log file).

Figure 2-10

Printing Error Logs Click Print to print the entire log file to your default printer.

---

Note - These logs can be very long. Make sure you want the entire file before printing it.

---

2.2.4.1 Schedule Window Button

Click the Schedule window button to display a pop-up menu like the one shown in Figure 2-11. The test schedule menu allows you to set start and stop dates and times to automatically run the SunDiag exerciser.

Figure 2-11

· How to Use SunDiag's Schedule Feature

1. Click Enable.

   This setting needs to be enabled for the schedule feature to work.

2. Fill in the necessary times and dates for automatic testing.

   Type times in the form of hours:minutes:seconds.

   For example: 2:30:00 runs SunDiag for two-and-a-half hours.

   Type dates in the form of month/day/year.

   For example: 1/15/93 represents January 15, 1993.

   Include the colons and slashes.

---

Note - You don't have to fill in all of the fields for the schedule feature to work. Stop Date, Stop Time, and Run Time override each other. If you set these fields with conflicting times, the SunDiag software will stop testing at the earliest specified time.

---

3. Click Apply.

   The SunDiag exerciser will begin automatic testing at the start time and continue testing until either the Run Time or Stop Time parameter is met.

2.2.4.2 Set Options Window Button

The Set Options menu sets various global test options, such as whether core files are enabled or disabled, and whether the tests should run on an error condition. Click SELECT on the Set Options window button to display the Set Options menu, like the one shown in Figure 2-12.

Figure 2-12

You can only change items on this menu when testing is stopped. The settings are enabled or disabled by clicking SELECT. Selected items appear in bold frames on monochrome monitors, and "pushed-in" on color monitors.

Core Files When you click Disable, the tests try to capture the signals that cause core dumps. This means that you receive informative messages instead of a core file. These messages will appear on the SunDiag console window, and they will be stored in the information log file /var/adm/sundiaglog/sundiag.info and in the error logs file /var/adm/sundiaglog/sundiag.err. See "Log Files Window Button" on page 2-11 for more details.

Single Pass When enabled, all SunDiag tests will only execute one time.

Quick Test When enabled, abbreviated versions of the enabled tests are run. See the individual test descriptions in Part 2 of this book for details; some tests perform normally when this option is specified.

Verbose When enabled, more diagnostic messages are printed in the console window while tests are running.

Trace When enabled, the test programs are traced down by function names. Messages are printed in the console window while the tests are running.

---

Caution - Enabling trace is not recommended for the normal operation of the SunDiag system exerciser. Trace should be used for debugging purposes only, and it should be started from the command line.

---

Auto Start To eliminate the need to click the Start button to start SunDiag testing:

1. Click the Auto Start setting.

2. Save the global options and test options (using the Option Files menu described in the next section) to a specified option file.

3. Specify the saved option file (using the -o argument) when you type the command line to invoke the SunDiag system exerciser. The program will start testing after the window is displayed.

Run On Error When enabled, SunDiag runs failing tests continuously until the number specified for Max. # of Errors is reached.

Send Email This feature is designed to keep you apprised of testing status by sending the most recent 60 lines of the SunDiag information file, /var/adm/sundiag.info. Email is sent to the host machine.

To select the automatic send mail feature, press MENU on the Send Email abbreviated menu button. The choices are:

• Disable..- no email is sent.
• On_Error - email is sent when the SunDiag software finds an error.
• Periodically - used in conjunction with the Log/Email Period option. When running, the SunDiag software sends email once every x minutes, where x is the period set in the Log/Email field.
• Both....- SunDiag sends email both on error and periodically.

Max. # of Passes If you enable this option, the SunDiag exerciser runs passing tests continuously until the number specified for Max. # of Passes is reached. Setting this option to 0 runs the tests indefinitely until interrupted.

Max. # of Errors See "Run On Error" above. A setting of 0 will continue testing no matter how many errors are encountered.

Concurrent Tests # Type the number of tests you want to run concurrently.

---

Note - If you set this parameter to a high value, then you may need to set the maxusers parameter to accommodate the increased number of UNIX processes. See Section 1.2.6, "Setting the Maximum Number of Processes," on page 1-8, for more information.

---

Log/Email Period This option sets the time period you want the email error message to be sent (if enabled) and enables the periodic status log. To set the Log/Email Period, type the number of minutes. A zero entry disables this function.

Email Address This option tells SunDiag where to send the email message. The default is root. To change the address, remove the root entry and type a new address.

Printer Name Type the name of the printer you want to use. The name should match an entry in your /etc/printcap file, such as lp1, or the name of the host to which the printer is attached. If you do not specify a printer, SunDiag uses the printer specified with the PRINTER environment variable. If there is no PRINTER variable, SunDiag prints on the printer specified by the lp entry in your /etc/printcap file. Any printer you specify must exist in the printcap file.

Instances Per Test See Section 4.1, "Scalability Options" for instructions.

PAM (Scalable) See Section 4.1, "Scalability Options" for instructions.

PAM (Non-scalable) See Section 4.1, "Scalability Options" for instructions.

The Reset and Apply buttons Click the Reset button to reset the options to the default values (shown in Figure 2-12) and close the window. The Apply button applies the changed options to the SunDiag tests when they are started. You must click Apply for the changes to take effect.

---

Note - "Unpinning" the Set Options window before clicking Reset or Apply negates any changes you make.

---

2.2.4.3 Option Files Window Button

Clicking the Option Files window button displays a pop-up window menu for the option files that exist in the /var/adm/sundiaglog/options directory. The options stored in these files are those you have chosen from the Set Options menu and from the control panel for specific tests. The Option Files button works only when tests are stopped.

Figure 2-13

· How to Save Options to an Option File

1. Type the name of your option file in the test field.

   The SunDiag default option file is named .sundiag. If you start the software without the -o option, the .sundiag file is used, if it exists.

2. Click Store.

   The SunDiag exerciser saves all options -- including those from the Set Options file and from individual test option menus -- to the specified file name.

· How to Load an Option File

1. Press MENU on the abbreviated menu button to select an option file.

   The options saved in the available option files listed here are the ones you set in the Set Options menu, in addition to the options chosen for each individual test.

2. Click Load.

   This loads your current settings from the named file. To write a new file, click Store, and then Click Load. The next time you start testing, the options you saved will be used, and the control panel will reflect these choices.

   Click Remove to remove the named option file. The SunDiag program will prompt you for confirmation before removing the file.

---

Note - In option files, the Processor Affinity filed is a binary value (or "bit position") that simulates check boxes in the OpenWindows software. See Section 3.7, "Setting Processor Affinity Masks in TTY Mode," on page 3-20 for more information about setting Processor Affinity Masks in ascii format.

---

2.2.4.4 Status View Button

In some large systems, such as the SPARCcenter^(TM) 2000 system, the SunDiag system exerciser may have as many as 100 disk tests running at one time. Each test is listed in the system status panel. You to quickly scroll the system status panel to a specific device using the Status View button.

The Errors Only option changes the system status panel to only list tests that have failed. Errors Only is the default option; clicking SELECT toggles between viewing Errors Only and All Tests. The first, last, previous, and next page items scroll the list to the desired page.

---

Note - The status view menu button is available if the system status panel is more than one page long, but certain choices (Next Page, Previous Page, First Page, and Last Page) are unavailable. Errors Only lists only tests that have failed. Go To Device has no effect.

---

Figure 2-14

2.2.4.5 Intervention Mode

This setting serves as a reminder that you must intervene before a test or tests can be run successfully. This setting affects two categories of tests:

• Tests of drives that require scratch media (tapes, discs, or diskettes)
• Tests that require loopback connectors

The SunDiag software will not run these tests until you have enabled intervention mode by clicking Enable for the "Intervention Mode" setting on the Control Panel. This setting does not change the program's behavior; it just serves as a reminder that you must intervene.

2.2.4.6 Test Setting

The test setting below the intervention setting is a quick way for you to select groups of tests. Click Default to enable the default group of tests. The lower part of the control panel will enable those tests. Click None to enable none of the individual tests. Click All to select all of the tests.

---

Note - Clicking SELECT on an individual test name ("disabling") while tests are running stops that particular test. Likewise, clicking None while tests are running is the same as clicking the Stop button.

---

2.3 Changing Individual Test Options

The lower part of the control panel lists the devices and corresponding tests available for testing. For example, fstest is a test that checks the file system structure for a disk, and so the SunDiag program provides an fstest for each installed disk drive. Options menu buttons for each particular test are also shown. Click the Options menu buttons to display menus to change individual test parameters. See "Options Menu Buttons" on page 2-26 for more details.

Devices are arranged into the following categories:

• MEMORY DEVICES
• DISK DEVICES
• CPU DEVICES
• SBUS DEVICES
• TAPE DEVICES

When started, the SunDiag software probes the kernel for available hardware devices and displays them in this part of the control panel.

When the SunDiag main window appears, the control panel shows all tests that have been enabled. This list includes all the devices found during the kernel probe, except those which require intervention or additional equipment. Clicking on the Start button will start all enabled tests.

Enable or disable tests by clicking the test names. Test names that are framed in bold or appear to be "pushed in" are selected. You can also enable or disable groups of tests by selecting the device category name.

If a device is enabled during testing, the test(s) are started as soon as possible. If an enabled device is disabled during testing, the test(s) stop immediately.

---

Note - tapetest may not stop until the tape has fully rewound.

---

2.3.1 Options Menu Buttons

All tests have options menus that allow you to specify test parameters. Options menus are displayed by clicking SELECT on the options window buttons to the right of each test listed on the control panel. (See Figure 2-2) All test options menus have configurations and options fields.

Figure 2-15 shows the rawtest option menu for SCSI disk #0. It shows that the disk has a 327.5 MB capacity, and that its controller ID is an SCSI CCS. This information comes from the operating system kernel. This test offers an exclusive setting control that can be "toggled" between read-only and read/write testing.

Figure 2-15 rawtest

Refer to the individual test descriptions in Part 2 -- "SunDiag Tests" for test-specific options. When you are ready to view another option menu, double-click on the pushpin to dismiss the option menu, or Reset or Apply any changes you have made. You can only view or make changes to one option menu at a time.

2.4 Running the SunDiag Program from an Icon

The SunDiag program continues to run if you close the SunDiag window to an icon. Close the SunDiag window as you would any other OPEN LOOK window by clicking SELECT on the window menu button on the upper left of the window frame. (You can also use the OPEN key keyboard shortcut). Even in icon mode, you can monitor the SunDiag program for test failures.

Figure 2-16