K-Class Troubleshooting

• Why do Extraneous Characters Appear in Front Panel Display Codes?
• What does it mean when "Memory not optimized" is displayed?
• How can I verify the firmware version (PDC code)?
  • Verifying Firmware Version with On-line Diagnostics
  • Verifying Firmware Version with Boot Console Handler (BCH)
• How do I troubleshoot & configure the internal modem?
  • General Comments
  • Internal or External
  • Software Configuration (HP-UX):
    • Dialin Script
    • Predictive Support/UX Modem Configuration
    • System Administration Manager (SAM)
    • Manual Configuration
    • Remote Console Setup
    • Remote Console Session
    • Troubleshooting
• Troubleshooting HP PowerTrust UPS configuration
  • What is an Uninterruptable Power Supply(UPS)?
  • What UPS products does HP have to offer?
  • What are the hardware installation and configuration requirements?
  • What are the software installation and configuration requirements?
  • What are the HP PowerTrust UPS commands (9.04 and 10.x)?
  • What are the HP PowerTrust configuration files?
  • What should I expect when powerfail occurs, and what should I expect when mains power returns?
  • Can I have more than one HP PowerTrust UPS on my system?
  • My Syslog file is filling up very rapidly with ups_mond messages, what is happening?
  • I have a SITE UPS can or should HP PowerTrust be connected to it?
  • How do I use Kermit to clear the UPS?
• What do the front panel display codes mean?
• Major Code "D0" first digit
  • 0  Reserved (not used)
  • 1
    • Interrupts
    • Selftests, Diagnostics
    • Boot Codes
  • 2   Cache (selftests and diagnostics)
  • 3   Processor Dependent Hardware (PDH)
  • 4   Late Selftests
  • 5   Bus Transactions
  • 6   Reserved
  • 7   Memory Subsystem Faults and PDCE_HPMC Memory Faults
  • 8   I/O Device Faults (K-Class/D-Class)
  • 9   Console Initialization Errors
  • A  Boot Device Initialization Errors
  • B  Operating System Panic (OS)
  • C 
    • System Initialization
    • HP-UX ISL Boot Codes
    • HP-UX System Initialization Codes
  • D  Shutdown Codes (OS)
  • E  Warning Codes (OS) (not used)
  • F  Run Codes (OS)
• What is an HPMC, and how do you access this information?

If your memory is not configured for optimum performance, the following warning message appears when you boot the system:

WARNING: Memory configuration is not optimized for performance. Refer to the System Installation or Memory Installation Manuals for memory configuration guidelines.

• Also, WARN 7704 appears on the LCD display when the message above is displayed.
• Your memory will still work, but your system's performance may be degraded.
• For optimum memory performance, memory must be installed in a particular slot sequence.
• Refer to the guidelines published at; K-class manuals (link) choose: System Memory Upgrade Guide
• A version of this problem can also be corrected by firmware version PDC 2.0 or greater. See Verifying CPU Firmware Version (PDC code) to verify and update your system.

Return to Contents

How can I verify the firmware version (PDC code)?

The current version of firmware can be determined using on-line diagnostics.  If on-line diagnostics are not present on your system, use the Boot Console Handler Procedure to verify the current revision.  Refer to the appropriate on-line diagnostic procedure to determine the version of firmware presently installed on your system.

Verifying Firmware Version with On-line Diagnostics

• For HP-UX systems prior to 10.20:
  1. At the system prompt, enter sysdiag
  2. At the DUI prompt, enter sysmap
  3. At the ENTER MAP prompt, enter cpumap
  4. Note the PDC Firmware value.
  5. At the ENTER MAP prompt, enter exit
  6. At the DUI prompt, enter exit

 

• For HP-UX systems 10.20 and later:
  (It is necessary to determine the Support Tools Manager version to determine which procedure to follow.)
  1. Login as root
  2. At the prompt, enter mstm
  3. Note the Support Tools Manager version.
  4. For Version A.07.00 and less, continue with Procedure A,  For Version A.08.00 and later, continue with Procedure B

 

• Procedure A: Support Tools Manager version A.07.00 and less:
  1. Press F2 (OK)
  2. Press F4 (device)
  3. Press F5 (select class)
  4. Move cursor to "Processor"
  5. Press the 'Enter' key on the keyboard
  6. Press F2 (OK)
  7. Press F8 (previous menu)
  8. Press F5 (tools)
  9. Press F2 (information)
  10. Press F2 (run)
  11. Press F3 (info log)   After a few seconds, the Information Tool Log for the CPU will be displayed.  The 'PDC Firmware Revision' is displayed in the first column.
  12. Note the PDC Firmware value.
  13. Press F2 (done)
  14. Press F8 (previous menu)
  15. Press F8 (previous menu)
  16. Press F8 (exit)
  17. Press F2 (OK)

 

• Procedure B: Support Tools Manager version is A.08.00 and later:
  1. Press F2 (OK)
  2. Move the cursor to "CPU"
  3. Press F3 (select)
  4. Press F4 (menu bar on)
  5. Enter "T" to select Tool Menu
  6. Enter "I" to select Information Mode
  7. Enter "R" to Run   After a few seconds, the Information Tool Log for the CPU will be displayed. The 'PDC Firmware Revision' is displayed in the first column.
  8. Note the PDC Firmware value
  9. Press F2 (done)
  10. Press F8 (exit)
  11. Press F2 (ok)

Return to Contents

---

Verifying Firmware Version with Boot Console Handler (BCH)

Boot Console Handler Procedure

1. Log on as root, and enter reboot -r
   This command will shutdown the operating system and reboot the system.
2. If AUTOBOOT is on, you will receive the following message:

Process is starting autoboot process.
To discontinue, press any key within 10 seconds.

3. At this point, press any key within 10 seconds to interrupt the booting process.
4. The Main Menu is displayed. The following prompt will appear:

MAIN MENU: Enter command or menu>

5. Enter the command: in
6. The following prompt will appear:

Information Menu: Enter Command>

7. Enter the command fv.
   The system will respond with the current firmware revision.

Creating the Firmware Update Tape

The firmware patch applicable for your system can be obtained through the HP IT Resource Center (ITRC). The procedure for creating a bootable tape with firmware, and the procedure for updating is included with the patch, after the patch file has been "unpacked'.
Unpack the patch file with the sh command.            

prompt: sh (patch name)

Two additional files will now be available. These files are xxxxyyyy.text and xxxxyyyy.frm.  The text file provides specific instructions for that release of firmware.  The frm file is the actual update image file.

Return to Contents

How Do I Troubleshoot & Configure the Internal Modem?

General Comments

The internal modem on the K-Class is provided so that the HP Predictive Support Software can alert your HP Customer Engineers of possible hardware events that need their attention. You should be aware of the fact that if you disable the modem or reconfigure it, your Predictive Software will not be able to use the modem and alert your HP CE of possible hardware related problems. With this in mind, you may want to consider adding an additional modem on another port for remote system administration activities. Most high speed modems can be used on other Modem (MDP or ADP) Mux ports and the Response Center can provide limited configuration information for many of them. Any use other than the intended purpose of the internal modem is not supported by the Hewlett Packard Response Center. This includes any getty process, kermit or UUCP.

• The internal modem is a Multitech 19.2kbps faxmodem.
• It is normally configured on the Remote Console port, which is /dev/tty0p7.
• The hardware address is: 10/4/0.7 (8/4/0.7 on the K100).
• Port 7 on the MDP mux will not work, as this is used by the modem, neither will port 0, which is the "CONSOLE".
• On some systems, port 1 may be in use for the "UPS"
• The MDP is an optional 8 port modem distribution panel which plugs into a connector on the back panel of the system.

Return to Contents

Internal or External Modem Configuration

The "K' series modem architecture provides for either an internal (always installed) or external modem configuration. The internal modem is installed on the system's multifunction card and the phone line is connected through a line adapter module (LAM). This module is country specific and is attached to the back of the multifunction card. The external modem can be connected through a DB25 RS-232 port also on the multifunction card. This port is chosen using one of the Access Port commands from the system console.

The procedure is as follows:

1. Turn the front panel keyswitch right from vertical position to the service position.
2. From the Console type [Ctrl] B to initiate the CM> prompt.
3. At the CM> prompt, enter ca (Configure system remote support modem port.)
4. The response will be:

Current terminal mode identification: HP mode terminal
Do you wish to change it? (Y[N]):

5. Enter y
6. The response will be:

Select new terminal mode: (0=HP, 1=NON_HP, 2=AUTO_PWON)

7. Enter 0 (This is typically set to 0 (HP) or 2 (AUTO))
8. The response will be:

The terminal mode is now: HP mode terminal
                -------------------------
Current remote console configuration: (current setting)
Do you wish to change it? (Y[N]):

9. Enter y
10. The response will be:

Select the new configuration (0=Internal Modem, 1=External modem, 2=AUTO):

11. Enter 2
12. The response will be:

The new selection is now: AUTO Identification at power on.

(Internal modem is currently selected)

NOTE: Auto is the default setting for this configuration. It will always select the internal modem if it is available and the Line Adapter Module is installed. If you wish to force the selection of the external modem you must choose option 1.
                -------------------------
Current remote support modem port configuration:
System identification: (current setting)
Do you wish to change the configuration? (Y[N]):

13. Enter y
14. The response will be:

Current system identification:
New system identification (limited to 1 to 20 displayable characters or space for none): <chosen name>

New configuration: (takes effect at next remote connection):
System Identification: <chosen name>

CM>

(Configuration complete)

15. At the CM> prompt, enter co (Takes you back into console mode)

You may now put the front panel keyswitch into the vertical (ON) position.

Return to Contents

SOFTWARE CONFIGURATION (HP-UX)

The operating system configuration for the modem can be accomplished in one of the following four ways: 

• Dialin Script
• Predictive Support/UX Modem Configuration
• SAM
• Manual Configuration.

DIALIN SCRIPT

Included with the operating system on the "K" series system is a program called DIALIN. This program gives you the ability to configure and disable the internal modem port. The syntax is as follows:

dialin [-b baud] action
where action = enable | disable | status

The way this works is when you type the command dialin enable the script will create the device files necessary to support the modem (ttyd0p7, cul0p7, and cua0p7) When you type dialin disable it removes the files.

Do not use the -b option on the dialup script.  The communication speed at this level should always be set to 19200. The modem will speed sense to whatever modem you connect with (in or out), but it will talk to the "K" series backplane at 19200 baud.

If you are running Predictive Support, the dialin script will not work. You will get the following message:

HP Predictive Support/UX is configured to run on this system. Use HP Predictive Support/UX to enable/disable dialin access.

If you are trying to configure the modem using the dialin command and get this message, the modem should already be configured and functional. If not then there is a problem with the Predictive configuration.

Return to Contents

PREDICTIVE SUPPORT/UX MODEM CONFIGURATION

To execute the predictive configuration program (psconfig) and enable or disable the modem proceed as follows:

/opt/pred/bin/psconfig (run the program)

1. At the main menu select option 4 - CONFIGURATION Options Menu
2. At the "Configuration Options" menu select option 2 - Transfer Information
3. At the "Configure Transfer Information" menu select option 2 SETUP modem port

The system will scan the hardware for currently configured modem devices and list the configuration. Once the scan is complete, the output will look like this:

1 - Callout device file [cul0p7] type [MUX]
2 - Hardware Address [10/4/0] port [7]
3 - Dialer name [ps_mt1932_tone]
4 - Automatic transfer [y]
5 - CCITT modem [n]
6 - Maximum modem speed [2400]
7 - Callin enabled [y]

Set up the modem ports with these values? (Type y, n, or // to cancel):n
Type a number or // to cancel:
Select number 7 (Callin enabled)
Type y or n for callin on /dev/ttyd0p7 or // to cancel:

If the modem dialin is enabled and you want to disable it, answer no to this question. It will remove the entry from inittab and disable the SESSION dialin.

To re-enable SESSION mode retrace your steps and reply yes to the dialin question. When remote console is enabled you will be able to connect regardless of the SESSION Status. (See remote console below)

Return to Contents

System Administration Manager (SAM)

1. Log on as root
2. Type SAM
3. Highlight Peripheral devices and press [RETURN]
4. Highlight Terminals and Modems and press [RETURN]
5. At the Terminals and Modems screen, press [TAB] to go to the menu bar.
6. Use the arrow and [RETURN] keys to select the Actions menu.
7. Use the down arrow key to highlight Add Modem and press [RETURN].
8. In the Add Modem screen, set or select the following parameters:
   1. The hardware path to the serial interface to be used by this modem.
   2. The port number to be used by this modem.(This should be port 7 )
   3. The speed (baud rate) to be used by this modem (19200 is the setting for the internal modem). (This is the serial line to the modem baud rate,Not the modem bps on-line rate.)
   4. Whether or not the modem will be used for calling out from your system.
   5. Whether or not the modem will receive incoming calls.
   6. Whether or not this is a CCITT Modem (European standard modem).Press [RETURN] to turn the CCIT modem selection on/off.  When you finish setting parameters,activate OK by pressing [F5] or tabbing to ok and pressing  [RETURN].
9. You will receive a series of messages indicating the progress of the task. When you receive the message: 

This modem has been added

Select OK

10. Exit the SAM by pressing EXIT[F8] then EXIT SAM[F8] or by tabbing to the file menu and selecting the first EXIT then EXIT SAM.

    Verify the device files with: 

    # lssf /dev/*0p7
    mux2 card instance 0 port 7 callout at address 10/4/0 /dev/cua0p7
    mux2 card instance 0 port 7 callout at address 10/4/0 /dev/cul0p7
    mux2 card instance 0 port 7 hardwired at address 10/4/0 /dev/tty0p7
    mux2 card instance 0 port 7 callin at address 10/4/0 /dev/ttyd0p7

    NOTE: The hardware path on the K100 will be 8/4/0

Return to Contents

MANUAL CONFIGURATION

NOTE: The external modem port is /dev/tty0p7. The modem device files are named the same as they were in HPUX 9.04, except of course, there is no "LU" number - but there is an Instance number. The major number will  vary and the minor number has been rearranged.

/usr/sbin/mksf -d mux2 -H 8/4/0 -p7 -f -v -a 0 (tty = /dev/tty0p7)
/usr/sbin/mksf -d mux2 -H 8/4/0 -p7 -f -v -a 1 (dialout = /dev/cul0p7, /dev/cua0p7)
/usr/sbin/mksf -d mux2 -H 8/4/0 -p7 -f -v -a 2 (dialin = /dev/ttyd0p7)

Notice the -f option adds the "H/W flow control bit" and the -a is for the "access-mode Port" (access modes are 0,1, or 2). The default access mode is 0 (Direct connect). The access-mode meanings are:

Access-Mode	Port Operation
0	Direct connect
1	Dial out modem
2	Dial in modem

Device file Examples:

# ll /dev/*0p7
crw-rw-rw- 1 bin bin 193 0x000701 Jun 14 11:16 /dev/cua0p7
crw-rw-rw- 1 bin bin 193 0x000701 Jun 22 12:48 /dev/cul0p7
crw--w--w- 1 bin bin 193 0x000700 Jun 1 11:14 /dev/tty0p7
crw--w---- 1 uucp bin 193 0x000702 Jun 27 08:58 /dev/ttyd0p7

# lssf /dev/*0p7
mux2 card instance 0 port 7 callout at address 10/4/0 /dev/cua0p7
mux2 card instance 0 port 7 callout at address 10/4/0 /dev/cul0p7
mux2 card instance 0 port 7 hardwired at address 10/4/0 /dev/tty0p7
mux2 card instance 0 port 7 callin at address 10/4/0 /dev/ttyd0p7

If you had used the "-f" option, then the hardware flow control bit would
be set and the minor number would have a "1" in the 5th position...

crw-rw-rw- 1 bin bin 193 0x000711 Jun 22 12:48 /dev/cul0p7
^
Here!
======================================================================

You still need a proper getty definition, which sets the protocol that the modem will use (internal or external modem). First edit the /etc/gettydefs file and add an entry similar to the following, at the BOTTOM of that file. You may want to make a backup copy of your gettydefs file first. You must use the "vi" visual editor for this operation, because the file will not work with tabs at the end of a line.

 

# This is a comment line. Make sure you have one and only one blank line
# before and after each definition in the gettydefs file. This is a sample
# for the internal modem (19.2 kbs faxmodem; setup at 19,200 baud)

19200M # B19200 SANE HUPCL CS8 ISTRIP IXANY TAB3
# B19200 SANE HUPCL CS8 ISTRIP IXANY TAB3
# login:
# 19200M

# Notice the *one* blank line after the definition.....
#

===================================================================

Now, test your new gettydefs file entry with this command:

/etc/getty -c /etc/gettydefs | more

If you get any error message, you will have to edit the file until it tests (or "parses") correctly. Now you can change the /etc/getty command in the /etc/inittab file to point to the correct getty definition:

a1:3:respawn:/usr/sbin/getty -h ttyd0p7 19200M # This is typical a0:3:respawn:/usr/lbin/uucp/uugetty -r -t 60 ttyd0p7 19200M # Typical w/ uucp

NOTE: 19200M is the speed or /etc/gettydefs entry, not the baud rate.  The /etc/gettydefs file sets protocols and baud rates for the login attempt and the user shell. It includes more than just the baud rate (ie. parity, flow control, data bits, etc.)

See manpages for inittab, stty, or gettydefs for more details.

Return to Contents

REMOTE CONSOLE SETUP

To enable remote console on the "K" series system,   do the following from the local console:

[CTRL]B
CM> er (enable remote)

The following will be displayed:

Current remote console access configuration:

Remote CS: Disabled
Mode: Multiple
Password:
Password faults:

Do you wish to change the configuration (Y[N]) (Enter Y)

Remote console access is now enabled

To change the parameters reply yes to the change request.

• Mode: This can be set to single or multiple.
  • In single mode the system will allow you to connect to the remote console once then revert to the session mode.
  • In multiple you need to turn off the remote console to secure this option.
• Password: You can set a password of up to 20 characters.
• Password Faults: You can set the number of password attempts before the modem will shutdown.

To disable remote console access proceed as follows:

[CTRL]B
CM> dr (Remote console access disabled.)

Return to Contents

REMOTE CONSOLE SESSION

To dial in using the remote console you would connect with the modem as with  any normal dialup. When the modems connect you will see the following:

@
Remote console enabled

To gain access to the console you must send a break from the remote terminal. For the operator to get access back while the session is still connected send a break from the console. While the remote console is active you have full control of the system to reboot, run offline diagnostics, and evaluate the system using the Access Port commands. Also, when you are using the remote console, all work will be echoed on the local console and when the local console is used, information will be echoed on the remote.

When you want to disconnect from the remote console type:

[CTRL]B
CM> di

This will disconnect the remote console and return the control to the local console. You must still disable remote console ([CTRL]B CM>dr) to secure the system.

Return to Contents

TROUBLESHOOTING

One of the first things to check is the Phone Line that is connected to the modem. This can be done by disconnecting the Telephone line from the back of the modem and connecting it to a known good telephone.If this is ok then reconnect the phone line to the back of the modem.

To unhang the modem port:

1. First follow this procedure:   

1. Logon as root
2. Run SAM
3. Select peripheral Devices
4. Select TERM & Modems
5. Highlight modem port in question (with space bar)
6. Select actions (use F4 softkey and tab to Actions)
7. Remove device
8. Go back to actions
9. Add MODEM
10. Add modem for port in question (10/4/0.7 or 8/4/0.7 on K100)
11. Exit SAM

2. At prompt, #cd /var/spool/uucp and remove any LCKxxxxx files associated with the port in question

3. At root, run kermit, then enter the
       C-Kermit> set line /dev/tty0p7
       C-Kermit> set speed 19200, Then the following message will be displayed:
       /dev/tty0p7, 19200 bps (This is what should be returned)
   Then enter the following:
       C-Kermit> connect

This is what the system should respond with:
    Connecting to /dev/tty0p7, speed 19200
    The escape character is Ctrl-\ (Ascii 28, FS)

Type the escape character followed by C to get back, or followed by ? to see other options.

WARNING:  It is very important that the instructions below are followed to the letter.  If you make any mistakes, exit Kermit and start this step over.

Type ATZ   (This tells the modem to break whatever its doing)

Type OK    (This is what should come back from the modem)

NOTE: (Hold the [CTRL] key down and press the [\] key -- let go of both keys and press the [C] key)

[Ctrl \C] (To return to C-Kermit prompt)

4. If the modem does NOT respond back with OK after connecting to the device file then go to STEP 6. Otherwise, continue with step 5.

5. This port should now be unhung if the software or if a process was locked on to the port. Try using the port at this time to see if the problem has cleared up. If the problem is resolved, go to step 8, otherwise go to step 6.

6. If the port still does not respond, then the only option is to HALT the system and cycle power on the system. It is very important that you HALT and not reboot.  A reboot or shutdown -r will not reset the internal modem, only a power cycle will.  You must do a shutdown -h. While the power is down, check the connection from the Core I/O Board to the Line Access Module (LAM).  This is the small module into which the phone line plugs.

7. If halting the system and recycling power do not work, then a service call should be logged through the ITRC call logging function or call your local HP hardware service number.

8. If the software and modem begin working after halting and recycling the power, make sure all current modem patches have been applied to resolve any potential problems in the future.

Verifying Modem Internal Settings:
    Log on as root and run kermit
    C-Kermit> set line /dev/cua0p7
    C-Kermit> set speed 19200
    C_Kermit> parity none
    C-Kermit> show (to confirm 8, parity none)
    C-Kermit> connect

There will be no prompt at this point, you are connected to the modem.

Type atl5

A string of characters will be returned. If &W0 and &E7 are in the string,  Then the modem has the correct values. Exit Kermit by typing:

[Ctrl\C] (Control-backslash c), then exit.

If instead &W1 and &E6 are returned then proceed to change the values as follows:

Type one line at a time and wait for OK to appear on the next line before proceeding.
    ATZ  
  AT   
  AT&E7
  AT&W8
  AT&W0

The modem parameters are now set. Type atl5 again to see the new parameters, verify &E7 and &W0 are set.  Exit Kermit by typing:

[Ctrl\C] (Control-Backslash C)

Return to Contents

Troubleshooting HP PowerTrust UPS configuration

What is an Uninterruptable Power Supply (UPS) ?

Return to Contents

---

What UPS Products does HP have to offer?

HP PowerTrust Product Line

PRODUCT NUMBER	DESCRIPTION
A2941A	600Va Standalone
A2994A	1300Va Standalone
A2996A	1.3KVa Rackmounted
A2997A	1.8KVa Rackmounted
A2998A	3.0KVa Rackmounted

The 1.3KVa and 1.8KVa devices feature a single removable battery box, while the 3.0KVa unit features two removable battery boxes.

Return to Contents

---

What are the hardware installation and configuration requirements?

Installation of these units must be performed by a qualified HP Customer Engineer(CE).

NOTE: The customer will be responsible for supplying necessary mains supply and connection of the mains supply to the UPS. The HP CE can do the rest, and will connect the mains input cable from the system into the UPS output port.

Then if peripherals are also to be protected by this UPS then the standard Power Distribution Unit (PDU) in the rack will be replaced by a specialist UPS PDU (unless the UPS was factory fitted in the rack) and all peripherals can then be connected to this PDU.

The only other connection is the UPSs serial RS-232 cable to a suitable UPS port or mux port on the host system.

It is important that either have a large enough single UPS unit to cover your load requirements, or that you purchase more than one UPS unit and split the equipment across the multiple UPS controllers.

In the event of UPS controller or battery problems - normally signified by the lights on the front panel of the UPS controller. The first level of defense is that the UPS has a power source switch on the rear of the controller that can be in two positions, Normal and Bypass.

• Normal - Normal UPS operation, mains power runs thru the UPS and is supplied to the host.
• Bypass - All UPS hardware is bypassed and this is like a standard mains supply.

Return to Contents

---

What are the software installation and configuration requirements?

The HP PowerTrust software is bundled with HP-UX 10.01 and newer versions, so there are no other software products to purchase, and the engineer or end user has a choice of using SAM or manual methods for configuration of the UPS system.

Loading the software

The software is found on the core-OS media bundle as part of the standard OS. The fileset to look in to specify the software to load if it hasn't been loaded is as follows:

HP-UX 10.01

Software on 10.01 can be found and loaded using swinstall, again the software is on the core-os media, and is under the CORE-OS fileset and is also called UPS-TOOLS, the loading process will be the same as any other fileset loading operation within swinstall.

1. Find the fileset and highlight it.
2. Either press [m] or use the mark for install option from the ACTIONS pull down menu.
3. Select the INSTALL ANALYSIS option from the ACTIONS pull down menu.

Configuring the Software

The software may be configured either using manual commands or by using SAM, either way is permissible and relatively straight forward.

Configuration through SAM

• Start SAM
• Select 'Peripheral devices' option from main menu
• Select 'Uninterruptable Power Supplies' option from this menu

You will now be presented with the following configuration screen:

Global Configuration

Daemon status:  Active

Configuration file:   /etc/ups_conf

Shutdown delay (Minutes):  1

Shutdown timeout (Minutes):  5

 

UPS Type	Device File	Hardware Path	Port Number	Shutdown Enabled
SOLA	/dev/tty0p	n nn.nn	n	Yes

Where n should replaced by the appropriate number for device, hardware path and port number.

All of the above fields can have their values modified, according to the configuration pertinent to your  system. Having changed the required fields, SAM will ask if you want to restart the /etc/ups_mond monitoring daemon in order for the changes to take effect.

 

Manual Configuration Method

The manual configuration method is as simple if not in some cases easier than using SAM. The above configuration fields can be found in the UPS configuration file /etc/ups_mond.

Completing the configuration and starting the daemon

The configuration is made complete by adding the ups_mond entry into /etc/inittab to ensure automatic startup at boot time - this entry is discussed in SECTION 6: CONFIGURATION FILES, so for the moment we need only know that the required ups_mond line must be added to the /etc/inittab file.

If the ups_mond daemon needs to be started by hand for any reason then we need to issue the following command:

HP-UX 10.01

/usr/bin/rtprio 0 /usr/lbin/ups_mond -f /etc/ups_conf

Return to Contents

---

What are the HP PowerTrust UPS Commands?

ups_mond(1m)

HP PowerTrust Uninterruptable power supply monitor daemon

When the daemon detects loss of ac power for a period of time exceeding a configured limit, ups_mond ensures file system integrity by shutting the system down in a graceful manner.

You can specify the -f option along with a path and filename to specify an alternate configuration file, other than /etc/ups_conf.

By default ups_mond is locked into memory and is not swappable, although it can be forced to swap if extreme caution is exercised. This caution takes the form of ensuring that all configured swap disks are protected by the UPS hardware so that in the event of a mains power failure, you do not lose the daemon process. Having ensured this the -s option can be specified to allow the daemon to be swapped.

The process is normally started from /etc/inittab and started by init(1m), when adding the ups_mond entry to /etc/inittab, you must ensure the entry follows the rc inittab entry as you must ensure that syslog is up and running before ups_mond starts.

ups_mond MUST always be started using rtprio(1) to set real time priority.

Messages from ups_mond are displayed on the system console and recorded in /var/adm/syslog/syslog.log(10.01)

power_onoff(1m)

power_onoff instructs the UPS monitor (ups_mond(1m)) to shut down the system, and optionally tells it when to power the system back up. The daemon then instructs the UPS hardware when to turn the power on and off. The time to restart the system(power on) is specified with power_onoff(1m) command line arguments.

The PowerTrust UPS hardware has a maximum limit on the elapsed time between power off and the power on time of 99.9 hours, if you set the options to exceed this value it will default back to 99.9 anyway.

Options

-n

No power on. This will cause the UPS to switch off and NOT be powered back on. A manual reset of the power WILL be required.

time

Can be specified as one, two or four digits to specify time - the following are allowable formats: 0815, 8:15, 8'15, 8h15, 8.15, 8,15

date

Can be specified as the day of the week spelt out, or a date as a day month and optional two digit year field. You can also specify the special names 'today' and 'tomorrow'. If no date is specified the command will assume today.

+increment

This parameter can be set after the date and time parameters to specify a further delay minutes, hours or days ahead.

For fuller details and command line examples review the power_onoff(1m) man page.

 

Where are the commands?

HP-UX 10.01

/usr/lbin/ups_mond(1m)

/usr/sbin/power_onoff(1m)

Return to Contents

---

What are the HP PowerTrust Configuration files?

There are two configuration files for HP PowerTrust UPS systems, they are /etc/inittab and /etc/ups_conf. This is an overview of the contents of both, and the valid settings.

/etc/inittab

ups::respawn:rtprio 0 /usr/lbin/ups_mond -f /etc/ups_conf (10.01)

Remember the UPS entry MUST follow the rc entry in /etc/inittab to ensure that syslogd has been started before ups_mond is started.

/etc/ups_conf

shutdown_delay_mins	1
shutdown_timeout_mins	5
upstty:/dev/tty0p1	# port 1 (second) on terminal mux 0 for UPS RS-232

Explanation of the above fields:

Shutdown_delay_mins:1(Default) (maximum 15 minutes)

This is the time the system waits to see if mains power will be restored, before issuing a shutdown -h 60 command.

Shutdown_timeout_mins:5(Default) (maximum 15 minutes)

This is the length of time in minutes that the UPS will supply power to the system after the shutdown -h 60 process has started.

WARNING: If this parameters value is set to be too short for your normal shutdown process to complete - UPS will stop power supply output to the system and your shutdown process will crash before completion. It is advised to time a standard shutdown process, in order that one can be reasonably certain of assessing the value for this parameter accurately.

upstty:/dev/tty0p1 (Default)

This is the field to be used to specify the port that your UPS serial connection is connected to, and down which it will communicate with the host system.

Shutdown Parameter Calculations

HP will only guarantee 15 minutes of battery life on the PowerTrust UPS system. The parameters shutdown_delay_mins and shutdown_timeout_mins MUST add up to no greater than 15 minutes.

So you can have 5 and 10, 3 and 12, 1 and 14 or 0 and 15 respectively; as long as the two added together DO NOT exceed 15 minutes. If you  try to exceed this 15 minute total limit for the total of both parameters, then ups_mond has the defaults (1 minute and 5 minutes) hard coded into it and it will ignore your values in /etc/ups_conf and revert back to the defaults.

WARNING:   As shutdown_delay_mins can be set to 15, this would mean that given the 60 second grace period issued as part of the shutdown command; once shutdown_delay_mins has expired, the shutdown will not start for 16 minutes. This together with the fact that the fastest shutdowns will take 5 minutes (best case),  would mean you need 21 minutes of battery life which you might not have. The problem occurs when ups_mond will signal withdrawal of power supply from the UPS at the 15 minute time limit, regardless of battery life.

Return to Contents

---

What should I expect when powerfail occurs, and what should I expect when mains returns?

First of all lets clarify that in UPS terms there are two types of powerfail event, short and long. Lets discuss the two types and then describe what happens in both cases.

SHORT

A short powerfail is quite simply a powerfail event that recovers before the value set for shutdown_delay_mins has expired - i.e. the mains power returns before we issue the shutdown -h 60.

LONG

A long powerfail event would be where mains power has not returned within the time value set for shutdown_delay_mins and we have issued the shutdown -h 60 command and are in the process of shutting down the system.

Events as they occur - Short powerfail

1. AC power is lost
2. UPS notifies its host system
3. Message written to console and syslog(9.04) or syslog.log(10.x)

Message on console reads:
ups_mond: UPS /dev/tty0p1 AC POWER FAILURE - running on UPS battery.
if power is not returned within previously configured time period,
your system will automatically go to graceful shutdown.

** Mains power returns within shutdown_delay_mins time value: 

1. Power restored
2. UPS notifies host system
3. Write message to console and syslog(9.04) or syslog.log(10.x)

Message on console reads:
ups_mond:UPS /dev/tty0p1 OK:AC power back on
ups_mond:AC power to all recognized system critical UPSs OK system will not shutdown

Events as they occur - Long powerfail

1. AC power is lost
2. UPS notifies its host system
3. Message written to console and syslog(9.04) or syslog.log(10.x)
   

   Message on console reads:
   ups_mond: UPS /dev/tty0p1 AC POWER FAILURE - running on UPS battery.
   if power is not returned within previously configured time period,
   your system will automatically go to graceful shutdown.

   ** Mains power has not returned **

4. shutdown_delay_mins expires - ups_mond is now using shutdown_timeout_mins and will signal UPS hardware to withdraw power in shutdown_timeout_mins 'minutes'.
5. shutdown -h 60 - command issued
6. shutdown_timeout_mins expires - UPS turns power off

If mains power returned after shutdown_delay_mins expired, UPS will NOT come back on line. it will remain in battery backup mode until shutdown_timeout_mins expires and the UPS turns off power. The UPS will stay powered off for a number of seconds and then turn back on.  Once the mains supply is re-established the system itself will regain power and start to boot.

Return to Contents

---

Can I have more than one HP PowerTrust UPS on my system?

Yes, the configuration remains exactly the same, there will still only be one entry for ups_mond in /etc/inittab, but:

1. The extra UPSs must be plugged into a tty port.
2. That tty reference must be added into an 'upstty' statement on a separate line in /etc/ups_conf.
3. And ups_mond restarted to pick up the new entries.

Return to Contents

---

My syslog file is filling up very rapidly with ups_mond messages, what is happening?

In this case a quick look in /etc/ups_conf to make a note of the tty ports being used, and then a comparison done against the tty ports set to respawn in /etc/inittab, should reveal that you have a getty respawning on the same port as the UPS connection is configured for.

In /etc/inittab either set the rogue getty entries to off or remove them completely, save the inittab file and from a ux command prompt type 'init q' to re-read the new inittab file - this should resolve the problem.

Return to Contents

---

I have a SITE UPS can or should HP PowerTrust be connected to it?

NO, PowerTrust requires sinisoidal input. Output from a site UPS is often clipped, which can cause PowerTrust to exhibit erratic behavior - such as rogue power failures and complete shutdowns.

Return to Contents

---

How do I use Kermit to clear the UPS ?

If you are having a problem where you are getting the following error message from ups_mond to the console, its cause is likely to be that the UPS is set to the wrong baud rate.

UPS /dev/tty2p1 Read Failed: File Exists: Uninterruptable Power Supply has not been connected correctly: Loss of power will not be detectable

The UPS should be set to 1200 baud. The UPS box can get into a bad state  The result is that the UPS is trying to communicate at the wrong baud rate or is not set up with any baud rate. To clear the UPS, do the following:

1. Make sure the ups_mond program is not running.
   1. Type ps -ef|grep UPS to see if it is.  If it is still running, copy the inittab file:  cp /etc/inittab /etc/inittab.backup 

2. Then edit the /etc/inittab file and look for a line like:
   ups1::respawn:rtprio 0 /etc/ups_mond -f /etc/ups_conf

3. Insert a # in front of line so it looks like
   #ups1::respawn:rtprio 0 /etc/ups_mond -f /etc/ups_conf

4. Save your changes. Then type: init q

5. Verify with ps -ef|grep ups that the /etc/ups_mond program is not running.

2. Look in /etc/ups_conf and check for port that this program uses. It is most likely /dev/tty0p1. Use this port for all subsequent commands where the word PORTNAME is shown.

3. Type kermit
   Prompt will be  C-Kermit>

4. A the C-Kermit> prompt type the following:
   set line PORTNAME <-----Remember to use name from above
   set speed 1200
   connect
   <-- no prompt. You type a carriage return If you get a ?? from ups, it is working fine and you just need to exit kermit - Skip to step 8.

5. If no ??, type [CTRL\C] <---- That is simultaneous CTRL and \ keys followed by a C. you will then get kermit prompt.  At the C-Kermit> prompt Type:
   set speed 9600
   connect  <-- no prompt. You type a carriage return. 

6. If no ??, try 2400 or 4800 baud. If none of the different baud rates work, the problem is possibly a bad or wrong cable, device file or mux board problem, bad UPS, or some other process is using the port. Go to step 8 to exit kermit. This procedure will not work.

7. When you get the ??, type [N] <-----------NO Carriage return. You will see NM
   Type [CTRL\C] <---- That is simultaneous CTRL and \ keys followed by a C. you will then get kermit prompt.
   Type:
   C-Kermit>set speed 1200
   C-Kermit>connect <-- no prompt. You type a carriage return
   You should now see ?? You might see "CM" instead. If so, hit [ENTER]and you should now get ??

8. Type [CTRL\C] <---- That is simultaneous CTRL and \ keys followed by a C. you will then get kermit prompt.
   Type
   C-Kermit>exit
   $ <------------- shell prompt. May be #.

9. Copy the original inittab file back by typing cp /etc/inittab.backup /etc/inittab
   Type init q  Verify that the ups_mond program is running.