Fujitsu Computer Accessories B23Q8YA G 02 F User Manual

B23Q8YA-G-02-F

PRIMEPOWER

Hardware Platform Guide: Operation

for Solaris™ 10 Operating System

-- Start Here--

Preface

Solaris™ Operating System is described as Solaris in this manual.

Purpose

This manual provides notes necessary after installing the Solaris 10 on PRIMEPOWER.

Intended Readers

This manual is intended for system administrators who install Solaris™ Operating System (Solaris) and

operate it.

Readers are expected to have the basic knowledge of a system administrator.

Organization

This manual is organized as follows:

z Chapter 1 Notes after the Installation

This chapter describes post-installation notes.

z Chapter 2 Backup and Restore of System Disk (PRIMEPOWER 1/200/250/400/450/600/650/850)

This chapter explains the procedures for backup/restore of the disk (hereafter called, the system disk)

containing the Solaris software.

z Chapter 3 Backup and Restore of System Disk (PRIMEPOWER 800/900/1000/1500/2000/2500)

This chapter explains the procedures for backup/restore of the disk (hereafter called, the system disk)

containing the Solaris software.

z Chapter 4 Secondary LAN port on PRIMEPOWER250/450

This chapter describes how to use the secondary LAN (10/100/1000Mbps) port on PRIMEPOWER

250/450.

z Chapter 5 Precautions

This chapter explains precautions when using Solaris 10.

High Risk Activity

The Customer acknowledges and agrees that the Product is designed, developed and manufactured as

contemplated for general use, including without limitation, general office use, personal use, household use,

and ordinary industrial use, but is not designed, developed and manufactured as contemplated for use

accompanying fatal risks or dangers that, unless extremely high safety is secured, could lead directly to

death, personal injury, severe physical damage or other loss (hereinafter "High Safety Required Use"),

including without limitation, nuclear reaction control in nuclear facility, aircraft flight control, air traffic

control, mass transport control, medical life support system, missile launch control in weapon system. The

Customer, shall not use the Product without securing the sufficient safety required for the High Safety

Required Use. In addition, Fujitsu (or other affiliate's name) shall not be liable against the Customer and/or

any third party for any claims or damages arising in connection with the High Safety Required Use of the

Product.

Trademarks

Sun, Sun Microsystems, the Sun Logo, Solaris and all Solaris based marks and logos are trademarks or

registered trademarks of Sun Microsystems, Inc. in the U.S. and other countries, and are used under license.

Microsystems, Inc. in the United States and other countries.

Java and all Java-based trademarks and logos are trademarks or registered trademarks of Sun Microsystems,

Inc. in the United States and other countries.

Microsoft, MS-DOS and Windows are registered trademarks of Microsoft Corporation in the United States

and/or other countries.

All SPARC trademarks are used under license from SPARC International,Inc. and are trademarks or

registered trademarks of that company in the U.S. and other countries. Products with the SPARC trademark

are based on architecture developed by Sun Microsystems,Inc.

All other products or company names mentioned in this document are claimed as trademark and trade

names by their respective companies.

March, 2005

October 2009 Sixth Edition

Notes

z This manual may not be copied by any means without the express written permission

of FUJITSU LIMITED.

z FUJITSU LIMITED reserves the right to make changes to this manual without prior

notice.

Contents

Chapter 1 Notes after the Installation...................................................................................................................1

1.1

PRIMEPOWER - Specific Notes.....................................................................................................................................1

Modifications of Solaris OS (Patches, PTF, Program Local Fix).............................................................................1

Changing the date and time......................................................................................................................................1

Serial port.................................................................................................................................................................1

1.1.1

1.1.2

1.1.3

1.1.3.1 PRIMEPOWER800/1000/2000............................................................................................................................1

1.1.3.2 PRIMEPOWER900/1500/2500............................................................................................................................1

1.1.4

Gigabit Ethernet port on PRIMEPOWER250/450...................................................................................................2

Fibre Channel Cards (PW008FC3, PW008FC2 or GP7B8FC1)..............................................................................2

Time for booting Solaris ..........................................................................................................................................2

Error messages displayed after booting from Solaris DVD .....................................................................................2

Solaris OS - Notes and Bug Information..........................................................................................................................3

The order of instance numbers of optional cards .....................................................................................................3

Notes on Solaris OS operation .................................................................................................................................3

Notes on PCI cards...................................................................................................................................................5

Setting Up PRIMEPOWER MIB Information.........................................................................................................6

Other notes and bug information..............................................................................................................................7

Unsupported Features ......................................................................................................................................................7

1.1.5

1.1.6

1.1.7

1.2

1.2.1

1.2.2

1.2.3

1.2.4

1.2.5

1.3

Chapter 2 Backup and Restore of System Disk (PRIMEPOWER1/200/250/400/450/600/650/850).....................9

2.1

2.2

Types of Backup Procedures............................................................................................................................................9

Backup and Restore of File System Unit .......................................................................................................................10

Backup of File System Unit ...................................................................................................................................10

Restore of File System Unit ...................................................................................................................................11

Backup and Restore of Disk Unit...................................................................................................................................12

Backup of Disk Unit ..............................................................................................................................................12

Restore of Disk Unit ..............................................................................................................................................12

2.2.1

2.2.2

2.3

2.3.1

2.3.2

Chapter 3 Backup and Restore of System Disk (PRIMEPOWER800/900/1000/1500/2000/2500).....................13

3.1 Types of Backup Procedures..........................................................................................................................................13

Chapter 4 Secondary LAN port on PRIMEPOWER250/450 ...............................................................................15

4.1

4.2

4.3

Outline ...........................................................................................................................................................................15

Notes..............................................................................................................................................................................15

How to Set Configuration Information ..........................................................................................................................15

Environment Setting ..............................................................................................................................................15

4.3.1

4.3.1.1 Determining the Client Name of Target Driver..................................................................................................15

4.3.1.2 Defining Hostname ............................................................................................................................................16

4.3.1.3 Defining Netmask Value....................................................................................................................................16

4.3.2

Setting the Operation Mode ...................................................................................................................................16

4.3.2.1 fjgi.conf file........................................................................................................................................................16

4.3.2.1.1 Parameters List............................................................................................................................................16

4.3.2.1.2 How to Set Parameter..................................................................................................................................18

4.3.2.1.3 Caution ........................................................................................................................................................18

4.3.2.2 Setting a JumboFrame........................................................................................................................................18

4.3.2.3 To use ndd(1M) Utility ......................................................................................................................................19

4.3.2.4 Changing the setting of FCode...........................................................................................................................22

4.3.2.5 Network Install...................................................................................................................................................23

4.3.2.6 Configuring VLAN interface .............................................................................................................................24

4.3.2.6.1 IEEE 802.1Q TagVLAN .............................................................................................................................24

4.3.2.6.2 How to create the VLAN interface..............................................................................................................25

Messages........................................................................................................................................................................26

Using GigabitEthernet on a Cluster System...................................................................................................................28

4.4

4.5

4.5.1

4.5.2

4.5.3

Support for Cluster System....................................................................................................................................28

Cluster Environment Setup Procedure ...................................................................................................................29

Notes in Cluster Environment................................................................................................................................29

Chapter 5 Precautions...........................................................................................................................................31

5.1

5.2

Limitations and known bugs for Solaris.........................................................................................................................31

Specification and incompatibility for Solaris.................................................................................................................40

Chapter 1 Notes after the Installation

This chapter describes the post-installation notes.

1.1 PRIMEPOWER - Specific Notes

1.1.1 Modifications of Solaris OS (Patches, PTF, Program Local

Fix)

Only the modifications (patches, PTF, Program Local Fix, etc.) for PRIMEPOWER provided by FUJITSU

LIMITED can be applied to existing PRIMEPOWER systems running Solaris OS software.

Do not use patches provided by Sun Microsystems, Inc. on the World Wide Web or those provided by other

vendors. Also be careful that you do not apply modifications (patches, PTF, Program Local Fix, etc.)

intended for other systems (GP7000D, etc.) provided by FUJITSU LIMITED.

For information on Solaris OS modifications, contact a FUJITSU systems engineer.

1.1.2 Changing the date and time

PRIMEPOWER 1/200/250/400/450/600/650/850

Use the date command in single user mode if you need to change the date and time for testing purposes.

PRIMEPOWER 800/900/1000/1500/2000/2500

See "System Console Software User's Guide" for information on how to change the date and time.

1.1.3 Serial port

1.1.3.1 PRIMEPOWER800/1000/2000

z All serial ports on the system board are to be used for the System Management Console.

z Do not access the following devices relating to serial ports. Also do not change the setting of

applications using the following device files (such as a port monitor).

/dev/term/a, b, ...

/dev/cua/a, b, ...

/dev/term/fa000, fa001, ...

/dev/cua/fa000, fa001, ...

/dev/se_hdlc

/dev/se_hdlc0, /dev/se_hdlc1, ...

However, devices that are not used as the console path may be used, but only for the FJVTS and

SunVTS. The usable default devices for FJVTS and SunVTS are /dev/term/fa003, fa004, ... and

/dev/se_hdlc3, se_hdlc4, ...

1.1.3.2 PRIMEPOWER900/1500/2500

z The device name of the serial port will be as follows:

/dev/term/fa000, fa001, ...

/dev/cua/fa000, fa001, ...

/dev/se_hdlc0, se_hdlc1, ...

Because the number of this device name is allocated from the instance number of driver, this number

does not necessarily correspond to the slot number of the system board.

z Do not change the setting of port monitor, etc. to the serial port for the console. Because the physical

device of the serial port used for the console is the physical device path which are linked from logical

device under /dev/fjse directory, the logical devices under /dev/term and /dev/cua, and /dev/se_hdlcX

files, which link to above-mentioned physical device path, are devices used for the console.

z Other than the serial port for consoles, five serial ports can be used by the user. However, keep in

mind that it is necessary to stop beforehand the application that is using the serial port of the system

board or the PCI/disk BOX in order to disconnect the system board by Dynamic Reconfiguration.

z Only the devices that are not used as a console path, can be used for FJVTS and SunVTS.

1.1.4 Gigabit Ethernet port on PRIMEPOWER250/450

PRIMEPOWER250/450 has a Gigabit Ethernet port as standard equipment. For how to use the Gigabit

Ethernet port, refer to "Chapter 4 Secondary LAN port on PRIMEPOWER250/450."

1.1.5 Fibre Channel Cards (PW008FC3, PW008FC2 or GP7B8FC1)

Install the following versions of drivers:

FUJITSU PCI Fibre Channel 3.0 or later.

To install the driver, use the installer in the driver CD-ROM that is provided with the Fibre Channel Card. If

you don't have the driver CD-ROM mentioned above, download it from the following page.

http://www.fujitsu.com/support/computing/server/unix/driver/

Install the driver again after an update of Solaris 10.

For more detailed information, refer to "FUJITSU PCI Fibre Channel 3.0 Installation Guide" that is

provided with the driver CD-ROM or the page described above.

1.1.6 Time for booting Solaris

When booting Solaris 10 10/08 or later from Solaris CD/DVD or from the network via network instances

such as hme/qfe, it may take more time (approx. 20 minutes) than previous releases.

Example

{0} ok boot net

Boot device: /pci@80,4000/network@1,1 File and args:

* This interval may be longer than previous releases.

SunOS Release 5.10 Version Generic_xxxxxx-xx 64-bit

....

1.1.7 Error messages displayed after booting from Solaris DVD

If Enhanced Support Facility has been already installed on the system, the following error messages (max.

32 lines) may be displayed on the console at the first time of the boot from the system disk after the boot

from the Solaris CD/DVD which is Solaris 10 10/08 or later, but the error messages can be ignored safely.

Sep 11 09:50:00 xxxx FJSVmadm[1505]: pw450:FJSVmadm:A:-:OBP:Detected

hardware error occurred at Jul 16 08:36:15 JST

Also, the following messages are stored many times in XSCF log, too. The error messages can be ignored

safely.

SCF> show-error-logs <Return>

***** No.01/32

Date

= 2008/09/16 09:47:00

SysDamage = ALARM(STOP)

FaultDetect = OBP

FaultCode = 260000FF

DetailInfo = [BOOTB]Boot Block error

ReplaceParts= ---

Syslog msg = Detected hardware error

30C97830 260000FF 08071523 3615FFFF

5B424F4F 54425D42 6F6F7420 426C6F63

6B206572 726F7200 00000000 00000000

00000000 00000000 00000000 00000000

1.2 Solaris OS - Notes and Bug Information

1.2.1 The order of instance numbers of optional cards

If an additional option card is installed, the instance number on the card might be assigned out of order.

Additional installation by Dynamic Reconfiguration or PCI Hot-plug is also applicable but replacement by

Dynamic Reconfiguration or PCI Hot plug is not applicable.

If you need to assign the instance numbers orderly, follow these steps below.

Set the following in /etc/system.

set mtc_off=1

Reboot the system.

Install the optional card.

After the above setting, it may take longer time to reboot the system. If that affects the system, delete the

line and reboot the system.

1.2.2 Notes on Solaris OS operation

CD-ROM/DVD-ROM in UNIX file system (UFS) format

Do not use any CD-ROM/DVD-ROM in UNIX file system (UFS) format that is not produced for the Solaris

OS.

Even if the file system in such a CD-ROM / DVD-ROM is in UFS format, it's file construction may differ

from that of a Solaris system. If such a CD-ROM / DVD-ROM is used, serious problems, including a

system crash may occur.

Enter the OpenBoot environment

The following operations on a running system will enforce the system to move to Open Boot mode, which

might damage data.

z Connecting / Disconnecting the keyboard cable to / from the system.

z Sending a Break signal from the console terminal connected to the serial port.

z Turning on/off the power to the unit connected to the serial port of the system.

(This power on/off is treated as a Break signal.)

To prevent a system from entering the Open Boot mode, do the following:

PRIMEPOWER 200/400/600

z Set the MODE switch on the operation panel of the system to "SECURE" mode.

z Set OBP command (PROM command) to disable a Break signal. For more detailed information, refer

to "PRIMEPOWER USER'S MANUAL."

PRIMEPOWER 250/450/650/850/800/900/1000/1500/2000/2500

z Set the MODE switch on the operation panel of the system to "LOCK" mode.

z Set OBP command (PROM command) to disable a Break signal. For more detailed information, refer

to "PRIMEPOWER USER'S MANUAL."

PRIMEPOWER1

z Change the default keyboard abort sequence effect with kbd commnd.

# /usr/bin/kbd -a disable <Return>

z Change the /etc/default/kbd setting

Edit the /etc/default/kbd as follows:

KEYBOARD_ABORT=disable

Set keyboard defaults from the /etc/default/kbd file.

# kbd –i <Return>

z Set OBP command (PROM command) to disable a Break signal. For more detailed information, refer

to "PRIMEPOWER USER'S MANUAL."

Single-user mode

If you enter the system in single-user mode with the following commands, the run-level of the init process

displayed by the who -r command may be displayed incorrectly. In this case, the commands pkgadd,

patchadd, etc. may terminate.

/sbin/init s

/sbin/init S

/usr/sbin/shutdown [-y] [-gN] -is (N: numeral)

/usr/sbin/shutdown [-y] [-gN] -iS (N: numeral)

To avoid this problem follow the steps below.

Enter the OpenBoot environment.

# /usr/sbin/shutdown -y -g0 -i0 <Return>

When the ok prompt is displayed, boot the system with the following command.

ok boot -s <Return>

Transport protocols for NFS version 2/3

Although NFS version 2/3 can use either TCP or UDP as the transport protocol, it is recommended to use

TCP. If UDP is used on a poor quality network(in which packet loss occurs or transfer takes too long time

etc.), overall performance, stability and / or security may be degraded and further problems may occur.

Special NFS client settings may be necessary to use TCP for NFS. For more detailed information, refer to

the NFS client manual.

From Solaris 2.6 or later, NFS clients can be made to use the TCP protocol, by specifying "proto=tcp" as the

mount option. If the protocol to be used is not specified, mounting NFS using TCP is first attempted, If that

fails, mounting NFS using UDP is attempted.

How to eject a diskette

When you want to eject a diskette, follow the instructions described below. If not followed, data may be

damaged.

For more detailed information, refer to "Managing Removable Media" in "System Administration Guide."

Make sure the diskette is not being used.

Run the following command.

# eject floppy0 <Return>

Make sure the Access LED has turned off.

Press the Eject button on the diskette drive unit, then eject the diskette.

WARNING messages of se driver

When a serial port receives a lot of data, such as when a large amount of text is pasted on a console screen

connected to a serial port, the console message "seN: Buffer overrun" (N means an instance number of se

driver) will be displayed and the data will be lost.

This is because the user process cannot read data fast enough.

The size of data pasted on a console screen should be 1024 bytes or less at a time.

Speed of reception depends on each user process.

1.2.3 Notes on PCI cards

How to install PCI cards

To install a new PCI card, follow the steps described below.

Also, when you remove the PCI card or change the slot, do the following steps.

For PRIMEPOWER 1

Enter the OpenBoot environment.

# /usr/sbin/shutdown -y -g0 -i0 <Return>

Enter the following command to stop the system in the OpenBoot environment at the next system

boot.

ok setenv auto-boot? false <Return>

Turn off the power to the system.

ok power-off <Return>

Install the PCI card to the system. (For more detailed information, refer to the manual attached to

the card.)

After you turn on the power, the OK prompt will be displayed. Boot the system with the following

command.

ok setenv auto-boot? true <Return>

ok boot -r <Return>

Some of the PCI cards need driver software. If that is the case, use the following steps:

Add driver software. (For more detailed information, refer to the manual attached to the card.)

Enter the Open Boot Environment. Some of the manuals attached to the card may instruct "Reboot

your machine", which means to enter the Open Boot Environment with the following command.

# /usr/sbin/shutdown -y -g0 -i0 <Return>

When the OK prompt is displayed, boot the system with the following command.

ok boot -r <Return>

For PRIMEPOWER 200/400/600

If the system is running, turn off the power to the system with the following command:

# /usr/sbin/shutdown -y -g0 -i5 <Return>

Install the PCI card to the system. (For more detailed information, refer to the manual attached to

the card.)

Set the MODE switch on the operation panel to "MANUAL" mode and turn on the power.

After you turn on the power, the OK prompt will be displayed. Boot the system with the following

command.

ok boot -r <Return>

Return the MODE switch on the operation panel to "SECURE" mode.

Some of the PCI cards need driver software. If that is the case, use the following steps:

Add driver software. (For more detailed information, refer to the manual attached to the card.)

Enter the Open Boot Environment. Some of the manuals attached to the card may instruct "Reboot

your machine", which means to enter the Open Boot Environment with the following command.

# /usr/sbin/shutdown -y -g0 -i0 <Return>

When the OK prompt is displayed, boot the system with the following command.

ok boot -r <Return>

For PRIMEPOWER 250/450/650/850/800/900/1000/1500/2000/2500

If the system is running, turn off the power to the system with the following command:

# /usr/sbin/shutdown -y -g0 -i5 <Return>

Install the PCI card to the system. (For more detailed information, refer to the manual attached to

the card.)

Set the MODE switch on the operation panel to "UNLOCK" mode and turn on the power.

After you turn on the power, the OK prompt will be displayed. Boot the system with the following

command.

ok boot -r <Return>

Return the MODE switch on the operation panel to "LOCK" mode.

Some of the PCI cards need driver software. If that is the case, use the following steps:

Add driver software. (For more detailed information, refer to the manual attached to the card.)

Enter the Open Boot Environment. Some of the manuals attached to the card may instruct "Reboot

your machine", which means to enter the Open Boot Environment with the following command.

# /usr/sbin/shutdown -y -g0 -i0 <Return>

When the OK prompt is displayed, boot the system with the following command.

ok boot -r <Return>

On PRIMEPOWER450 (Rackmount Type(7U))/900/1500 (with SPARC64 V)/2500, you are able to use PCI

Hot Plug function, which is included in Enhanced Support Facility. If you use PCI Hot Plug function, refer

to "PCI Hot Plug User’s Guide" in Enhanced Support Facility.

1.2.4 Setting Up PRIMEPOWER MIB Information

If you introduce the network management software and want the machine to be monitored using its proper

model name, you have to set up the MIB information as described below.

Stop the Solstice Enterprise Agent (SEA) daemon.

# /etc/init.d/init.snmpdx stop <Return>

Check the model name with the following command. Note the displayed model name.

# /usr/sbin/FJSVmadm/hrdconf | head -1 <Return>

Example:

PRIMEPOWER400

Modify the /etc/snmp/conf/snmp.conf file.

Before modification:

sysdescr Sun SNMP Agent,

After modification:

sysdescr Fujitsu SNMP Agent, model_name

Write the model name displayed in the step 3, in the place of “model_name”.

You can customize the above sysdescr string the way you like. But some network management software

refers to "model-name", please check whether SNMP is running on the system.

Restart the Solstice Enterprise Agent (SEA) daemon.

# /etc/init.d/init.snmpdx start<Return>

1.2.5 Other notes and bug information

For other notes and bug information during the system operation, refer to the following document:

z "Solaris 10 Release Notes "

1.3 Unsupported Features

The following features described in Solaris OS manuals are not supported.

Table 1.1 Unsupported features

No. Description

Alternative method

None

Reference manual

The Suspend-Resume features

"Solaris Common Desktop

Environment: User's Guide" in

"Solaris 10 Common Desktop

Environment User Collection"

man pages section 1M: System

Administration Commands

cannot be used with

PRIMEPOWER.

The pcitool(1M) command cannot None

be used with PRIMEPOWER.

Note:

There are some differences between the power management system provided by Solaris OS and Auto

Power Control System provided by Enhanced Support Facility. In the Auto Power Control System, you

can set AutoShutdown and AutoWakeup features on PRIMEPOWER. However, Auto Power Control

System is not supported with PRIMEPOWER1.

Chapter 2 Backup and Restore of System Disk

(PRIMEPOWER1/200/250/400/450/600/65

0/850)

This chapter explains the procedures to backup/restore the disk (hereafter called, the system disk) containing

the Solaris OS software.

2.1 Types of Backup Procedures

Backup of the system disk requires that the disk and slices to be backed up not be in use. The system disk

contains file systems (such as the root (/) file system) that cannot be unmounted. In backing up the system

disk, special procedures are required.

The following are some of the typical ways to back up the Solaris OS disk.

Understanding them will help you choose the appropriate method to back up your system disk properly.

For more detailed information about the procedure mentioned in a., see each subsection from "2.2 Backup

and Restore of File System Unit."

For more detailed information about the procedure mentioned in d., see each subsection from "2.3 Backup

and Restore of Disk Unit."

To back up a file system unit, use the ufsdump(1M) command after booting in single-user mode.

If you are restoring, use the ufsrestore(1M) command after booting from the Solaris OS CD/DVD to

restore the file system unit.

The file system that will be backed up is only the ufs file system.

Disk partition information such as the size of each file system will not be backed up. You must

record this information separately.

The ufsdump(1M) command backs up a file system divided into file units so it is possible to

restore to a disk having a different file system size. If only a limited amount of space has been

used in the file system, this may back up into a space smaller than the size of the file system.

When backing up using the ufsdump(1M) command, it is possible to only restore selected files.

To back up a file system unit, use the ufsdump(1M) command after booting from the Solaris OS

CD/DVD.

If you are restoring, use the ufsrestore(1M) command after booting from the Solaris OS CD/DVD to

restore the file system unit.

The file system that will be backed up is only the ufs file system.

Disk partition information such as the size of each file system will not be backed up. Therefore,

you must record this information separately.

The ufsdump(1M) command backs up the file system divided into file units, so it is possible to

restore to a disk having a different file system size. If only a limited amount of space has been

used in the file system, this may back up into a space smaller than the size of the file system.

When backing up using the ufsdump(1M) command, it is possible to only restore selected files.

To back up a slice unit, use the dd(1M) command after booting from the Solaris OS CD/DVD.

If you are restoring, use the dd(1M) command after booting from the Solaris OS CD/DVD to

restore the slice unit.

If you back up slice 0, it will only be possible to restore to a disk having the same physical

properties (disk space, number of cylinders) as the original.

Disk partition information such as the size of each slice will not be backed up. You must record

this information separately.

To back up a disk unit, use the dd(1M) command after booting from the Solaris OS CD/DVD.

If you are restoring, use the dd(1M) command after booting from the Solaris OS CD/DVD, to

restore the disk unit.

Because this procedure backs up the disk unit, it is only possible to restore to a disk having the

same physical properties (disk space, number of cylinders) as the original.

Depending on the tape device, you might have to observe some precautions when backing up or restoring.

Consult the manual of the tape device before such actions.

It is recommended that you select the SCSI device having the target ID 0 for Solaris OS installation. In this

case, the device name of the system disk will be "c0t0d0" obtained from the controller number (0), target ID

(0), and LUN number (0). If you install Solaris OS on a disk, having a controller number other than 0, the

controller number may change later depending on whether booting took place from a CD/DVD or a hard

disk. Thus when referring to disk device, check that it's name haven't changed due to the reason described

above.

If you apply the optional software that contains modules that run as part of the driver and file system of the

system disk, carefully follow the instructions and notes described in the manuals supplied with each

software option.

For example, if you configure disk mirroring using PRIMECLUSTER GDS, you must select the backup

procedures mentioned in b. In addition, if the system disk contains a file system other than a ufs file system

such as PRIMECLUSTER GFS, you must select the backup procedures using the backup command

provided in the file system or the dd(1M) command.

2.2 Backup and Restore of File System Unit

The following describes how to back up the file system unit using the ufsdump(1M) command. "Solaris

System Administration Guide" also describes the following procedures.

Note that disk partition information such as the size of each slice will not be backed up. Therefore, you must

record the information separately using the prtvtoc(1M) or format(1M) commands.

# prtvtoc XXXX <Return>

# format XXXX <Return>

format> partition

partition> print

The "XXXX" in the above example is the character device name of the target disk. Use slice 2 of the disk,

like :

/dev/rdsk/c0t0d0s2.

2.2.1 Backup of File System Unit

In single-user mode use the following procedures to back-up each file system. After backing up, reboot the

system.

Boot the system in single-user mode.

ok boot -s <Return>

Backup the file system with the ufsdump(1M) command.

Example: Backup to the "/dev/rmt/0."

# ufsdump 0ucf /dev/rmt/0 / <Return>

2.2.2 Restore of File System Unit

When restoring the file system unit, the Solaris OS is booted from the Solaris 10 CD1 or Solaris 10 DVD.

After booting Solaris OS, mount the disk to be restored temporarily, and then restore the file system unit.

Follow the procedures below to restore each file system. After restoring, reboot the system. If using a new

disk or a disk previously used for other purposes, create a boot block after creating disk slices and a disk

label with the format (1M) command, before starting the restoration process.

If optional software like PRIMEPOWER GDS, which runs as a part of the kernel is installed in the system,

extra care should be taken as the Solaris CD/DVD does not contain these modules. In such a case, consult

the manual of the respective optional software.

Boot the Solaris OS from the CD/DVD in single-user mode.

ok boot cdrom -s <Return>

If you use a new disk or a disk previously used for other purposes, create a disk slices and a disk

level with the format(1M) command.

# format <Return>

For information on how to create a disk slice and a disk level using the format(1M) command, see the

manuals.

Create a boot block with the installboot(1M) command.

#installboot/usr/platform/`uname-i`/lib/fs/ufs/bootblkXXXX<Return>

For information on how to create a boot block using the installboot(1M) command, see the manuals.

The "XXXX" in the above example is the character device name of the target disk. Use slice 0 of the

disk, like : /dev/rdsk/c0t0d0s0.

Create a new file system with the newfs(1M) command.

# newfs XXXX <Return>

The "XXXX" in the above example is the character device name of the target disk, like :

/dev/rdsk/c0t0d0s0.

Mount the file system with the mount(1M) command.

# mount -F ufs YYYY /mnt <Return>

The "YYYY" in the above example is the block device name of the target disk, like : /dev/dsk/c0t0d0s0.

Change to the mounted directory.

# cd /mnt <Return>

Restore the file system with the ufsrestore(1M) command.

Example: Restore from the "/dev/rmt/0."

# ufsrestore rvf /dev/rmt/0 <Return>

Remove the restoresymtable file created by the ufsrestore(1M) command.

# rm restoresymtable <Return>

Change the directory to the root (/), and unmount the file system.

# cd / <Return>

# umount /mnt <Return>

Check the integrity of the file system with fsck(1M) command.

# fsck XXXX <Return>

The "XXXX" in the above example is the character device name of the target disk, like :

/dev/rdsk/c0t0d0s0.

2.3 Backup and Restore of Disk Unit

When backing up a disk unit, use the dd(1M) command. The following describes how to back up a disk unit

to a tape device using the dd(1M) command. The backup records physical properties such as disk space and

a number of cylinders, so it is only possible to restore to a disk having the same physical properties. The

prtvtoc(1M) command displays the physical properties as " Dimensions."

# prtvtoc XXXX <Return>

Specify the slice 2 of character special disk (/dev/rdsk/...) to be backed up such as "/dev/rdsk/c0t0d0s2"

with the format(1M) command.

2.3.1 Backup of Disk Unit

In the backup of a disk unit, the Solaris OS is booted from the Solaris 10 CD1 or Solaris 10 DVD. After

backing up, reboot the system.

Boot the Solaris OS from the CD in single-user mode.

ok boot cdrom -s <Return>

Back up the disk unit with the dd(1M) command.

Example: Backup to the "/dev/rmt/0."

# dd if=XXXX of=/dev/rmt/0 bs=64k <Return>

Specify the character special disk name (/dev/rdsk/c0t0d0s2) to be backed up such as "/dev/rdsk/..." with

the dd(1M) command. In this case, you must specify the "s2" as the entire disk.

2.3.2 Restore of Disk Unit

In the restore of a disk unit, the Solaris OS is booted from the Solaris 10 CD1 or Solaris 10 DVD. After

restoring, reboot the system.

Boot the Solaris OS from the CD in single-user mode.

ok boot cdrom -s <Return>

Restore the disk unit with the dd(1M) command.

Example: Restore from the "/dev/rmt/0."

# dd if=/dev/rmt/0 of=XXXX bs=64k <Return>

Specify the character special disk name (/dev/rdsk/c0t0d0s2) to be backed up such as "/dev/rdsk/..." with

the dd(1M) command. In this case, you must specify the "s2" as the entire disk.

Chapter 3 Backup and Restore of System Disk

(PRIMEPOWER800/900/1000/1500/2000/2

500)

This chapter explains the procedures to backup/restore the disk (hereafter called, the system disk) containing

the Solaris OS software.

3.1 Types of Backup Procedures

Backup of the system disk requires that the disk and slices being backed up not be in use. The system disk

contains file systems (such as the root (/) file system) that cannot be dismounted. In backing up the system

disk, the special procedures are required.

The following are some of the typical ways to back up a Solaris OS system disk. Understanding them will

help you choose the appropriate method to back up your system disk properly.

For detailed information about the backup procedures mentioned in b. and d., see " System Console

Software User's Guide."

To back up a file system unit, use the ufsdump(1M) command after booting in single-user mode.

If you are restoring, use the ufsrestore(1M) command after booting from the Solaris OS CD/DVD to

restore the file system unit.

The file system that will be backed up is only the ufs file system.

Disk partition information such as the size of each file system will not be backed up. You must

record the information separately.

The ufsdump(1M) command backs up the file system divided into the file unit, so it is possible

to restore to a disk having a different file system size. If a limited amount of space has been used

in the file system, this may back up into a space smaller than the size of the file system itself.

When backing up using the ufsdump(1M) command, it is possible to restore only the file that

you select.

To back up a file system unit, use the ufsdump(1M) command after booting from the Solaris OS

CD/DVD.

If you are restoring, use the ufsrestore(1M) command after booting from the Solaris OS CD/DVD to

restore the file system unit.

The file system that will be backed up is only the ufs file system.

Disk partition information such as the size of each file system will not be backed up. You must

record the information separately.

The ufsdump(1M) command backs up the file system divided into the file unit, so it is possible

to restore to a disk having a different file system size. If limited amount of space has been used

in the file system, this may back up into a space smaller than the size of the file system. In the

backup using the ufsdump(1M) command, it is possible to restore only the file that you select.

To back up a slice unit, use the dd(1M) command after booting from the Solaris OS CD/DVD.

If you are restoring, use the dd(1M) command after booting from the Solaris OS CD/DVD to

restore the slice unit.

If you back up slice 0, it will only be possible to restore to a disk having the same physical

properties (such as disk space and a number of cylinders) as the original.

Disk partition information such as the size of each slice will not be backed up. You must record

the information separately.

To back up the disk unit, use the dd(1M) command after booting from the Solaris OS CD/DVD.

If you are restoring, use the dd(1M) command after booting from the Solaris OS CD/DVD to

restore the disk unit.

The procedure divides the disk unit for back up, so it is only possible to restore to a disk having

the same physical properties (such as disk space and a number of cylinders) as the original.

It is recommended that you select the SCSI device with target ID 1 for Solaris OS installation. Before

selecting the disk to install Solaris OS, make sure it is connected to the SCSI adapter plugged into the PCI

slot of PCI0B. In this case, the device name of the system disk will be "c0t1d0" (obtained from the

controller number (0), target ID (1), and LUN number (0)). If you install Solaris OS to other than the disk

having the device name "c0t1d0", (especially, a disk having the controller number other than 0) take care to

carefully obtain the correct device name.

If applying the optional software that runs as part of the driver and file system from the system disk,

carefully follow the instructions and notes described in the manual of each option. For example, if you

configure disk mirroring using PRIMECLUSTER GDS, you must select the backup procedures described in

a. above, because the PRIMECLUSTER GDS driver does not exist in the CD/DVD files. In addition, if the

system disk contains a file system other than the ufs file system such as PRIMECLUSTER GFS, select the

backup procedures using the backup command provided in the file system or the dd(1M) command.

Chapter 4 Secondary LAN port on

PRIMEPOWER250/450

This chapter describes how to use the secondary LAN (10/100/1000Mbps) port on PRIMEPOWER250/450.

4.1 Outline

The secondary LAN port on PRIMEPOWER250/450 is controlled by FUJITSU PCI Gigabit Ethernet

Driver (Interface name is “fjgi”) included in Solaris 10.

4.2 Notes

z When using 1000 Mbps communication, be sure to set Autonegation to 'on'.

z If using the PCI card (PW008GE4/PW008GE5/PW008QG1), please refer to the User’s guide attached

with the PCI card.

4.3 How to Set Configuration Information

This section describes required setting to use a fjgi interface.

4.3.1 Environment Setting

This section describes how to edit files required to build environment.

4.3.1.1 Determining the Client Name of Target Driver

To use the secondary LAN port as an IPv4 interface:

To allocate the IP address or hostname to the fjgi interface, edit /etc/hostname.fjgi* (* is an numerical

value) file. About netmask value, please see "4.3.1.3 Defining Netmask Value" .

Example /etc/hostname.fjgi*:

Define the unique hostname:

giga-v4

or, define the IP address:

192.168.150.1

Note: If you defined the hostname in hostname.fjgi* file, see the section "4.3.1.2 Defining Hostname "

for detailed information.

To use the secondary LAN port as an IPv6 interface

To allocate the IP address and prefix to the fjgi interface, edit /etc/hostname6.fjgi* (* is an numerical

value) file.

Example /etc/hostname6.fjgi*:

Define the unique hostname (This sample named giga-v6):

addif giga-v6/120 up

Note: If you defined the hostname in hostname.fjgi* file, see the section "4.3.1.2 Defining Hostname"

for detailed information.

To use the GigabitEthernet as an VLAN interface:

See "4.3.2.6 Configuring VLAN interface", and after defining VLAN name, proceed with above 1 or 2

procedure.

4.3.1.2 Defining Hostname

Defining an IP address with a unique hostname.

To use secondary LAN port as an IPv4 interface:

Edit IPv4 address and hostname into the /etc/hosts file.

Example /etc/hosts file:

# IP Address

190.168.150.1

Hostname

giga-v4

To use secondary LAN port as an IPv6 interface:

Edit IPv6 address and hostname into the /etc/inet/ipnodes file.

Example /etc/inet/ipnodes file:

# IP Address

Hostname

fe80::2e0:ff:fea6:2222 giga-v6

4.3.1.3 Defining Netmask Value

To use secondary LAN port as an IPv4 interface, define a netmask value with the IP address into

/etc/netmasks file.

Example /etc/netmasks file:

# IP Address

192.168.150.0

netmask

255.255.255.0

4.3.2 Setting the Operation Mode

This section explains how to edit the file or implement a command to change the operation mode. When you

use a secondary LAN port without the default method of automatic setup of speed/duplex/flowcontrol by

autonegotiation, change the operation mode with the procedures described below:

4.3.2.1 fjgi.conf file

When not using the default operation mode described in 4.3.2, you should use the

/platform/sun4us/kernel/drv/fjgi.conf file. The set-up procedure of this file is described below:

4.3.2.1.1 Parameters List

The following table shows the list of the parameters in which setting change is possible.

Table 4.1 Parameters List

Parameter

Value

Description

name

fjgi

When using PRIMEPOWER250/450

secondary LAN port, please set

name="fjgi".

FJSV,pgtb

FJSV,pgsb

When using PCI card PW008GE5,

please set name="FJSV,pgtb".

When using PCI card PW008GE4,

please set name="FJSV,pgsb".

parent

Please refer to “4.3.2.1.2 How to Set

Parameter”.

The location of the device node in a

device tree is specified.

unit-address

AutoNegotiation_A

Please refer to “4.3.2.1.2 How to Set

Parameter”.

The address within a device node is

specified.

On (Default)

AutoNegotiation is enabled. (In this

setting, unable to connect to the

remote device of a fixed mode.)

When you use 1000Mbps, please use

in Autonegotiation_A="On" (default)

mode.

Off

AutoNegotiation is disabled. (Fixed

mode)

LinkSpeed_A

Auto (Default)

Connect in 1000/100/10Mbps by the

connection partner.(This can be set

when AutoNegotiation_A="On".)

Connect in 1000Mbps.

1000

100

Connect in 100Mbps.

Connect in 10Mbps.

DuplexCapabilities_A

Both (Default)

Indicating that both Full-Duplex and

Half-Duplex are being supported.

(This

can

set

when

AutoNegotiation_A="On".)

Half

Half-Duplex operation is allowed.

Full-Duplex operation is allowed.

A flow control is performed according

to a setup of partner equipment.

Full

FlowControl_A

Auto (Default)

Rem

Only the flow control from the remote

machine is allowed. The flow control

to the remote machine is not allowed.

Only the flow control from the main

unit is allowed. The flow control from

the remote machine is not allowed.

Flow control is disabled.

LocSend

None

Parameter

Role_A

Value

Description

Auto (Default)

Auto select Master or Slave by the

partner. (This can be set when

AutoNegotiation_A= On.)

Master

Slave

Communicate by Master.

Communicate by Slave.

4.3.2.1.2 How to Set Parameter

The method of setting up of each parameter of the secondary LAN port (FJSV,pwga) shown in Table 4.1 is

explained. If using PCI card (PW008GE4/PW008GE5/PW008QG1), please refer to the User’s guide

attached with PCI card.

Search /etc/path_to_inst file for a fjgi interface.

# grep fjgi /etc/path_to_inst <Return>

"/pci@83,4000/FJSV,pwga@1" 0 "fjgi"

The interface-related parameters should be written in the /platform/sun4us/kernel/drv/fjgi.conf file.

Example: 1000Mbps Full-Duplex is set.

name="fjgi" parent="/pci@83,4000" unit-address="1" AutoNegotiation_A="On"

LinkSpeed=”100” DuplexCapbilities_A="Full" FlowControl_A="None";

name:

In the above example, the driver name is "fjgi".

parent:

The left-hand portion of the pathname, containing the words '/FJSV,gwga@' in

/etc/path_to_inst, is to be added here.

In this example, parent value is "/pci@83,4000".

unit-address: Please refer to the /etc/path_to_inst, and write the number (right number of

"/FJSV,pwga@..". In this example, unit-address value is "1".

After 2, the fjgi.conf setting is reflected by reboot.

# /usr/sbin/shutdown -i6 -y -g0 <Return>

4.3.2.1.3 Caution

Setting by fjgi.conf is necessary only when autonegation is not used for speed/duplex/flow or

connecting partner cannot support autonegotiation.

Since a setup the operation mode by the fjgi.conf file becomes effective per physical interface,

when the VLAN interface is being used, a VLAN interface is also changed simultaneously.

4.3.2.2 Setting a JumboFrame

Use of JumboFrame makes it possible to expand the TCP/IP MTU size from 1500 to 9000. As the

per-packet transmission size is expanded, the number of packets processed becomes fewer and thus the CPU

load is reduced.

The procedure of setting JumboFrame is show below.

a. Add the following line to /etc/system file.

set fjgi:fjgi_jumbo=1

b. Create /etc/fjmtu.fjgiX file and write a MTU-size between 1500 to 9000. And reboot, TCP/IP

MTU-size will be set. If you don't create this file, TCP/IP MTU-size will be the default of 9000.

vi /etc/fjmtu.fjgiX

"X" is instance number. Instance number must be the same as /etc/hostname.fjgi* file's *.

Notes

If you set a JumboFrame, it will be unable to transmit and receive the data of IEEE802.3 form.

If you use two or more cards, or VLAN interfaces, enabling JumboFrame function becomes

effective to all cards and all interfaces.

When you perform file transfer by FTP, please set MTU size to less than 8232byte.

Please use Jumbo Frames function only in 1000 Mbps communication. If you use Jumbo Frames in

10 Mbps or 100 Mbps, you won't be able to use the Jumbo Frame function.

4.3.2.3 To use ndd(1M) Utility

By using the ndd command, mode of operation can be changed dynamically.

Below shows how to reference and set each parameter by ndd command with examples.

Reference:

ndd -get /dev/fjgi 'param'

(param: each parameter)

example) # ndd -get /dev/fjgi 'link_status'

Settings:

ndd -set /dev/fjgi 'param' value

value)

(param: each parameter, value: each

example1) # ndd -set /dev/fjgi 'instance' 2

example2) # ndd -set /dev/fjgi 'adv_autoneg_cap' 0

To reflect the changed setting:

ndd -set /dev/fjgi 'instance' value

(Set the relevant interface instance number.)

ndd -set /dev/fjgi 'adv_autoneg_cap' 0

(Once, turn off the function of AutoNegotiation.)

ndd -set /dev/fjgi 'adv_autoneg_cap' 1

(Then, turn on the function of AutoNegotiation.)

example) In case, to set the setting of fjgi0 1000Full => 100Half.

# ndd -set /dev/fjgi 'instance' 0

# ndd -set /dev/fjgi 'adv_1000fdx_cap' 0

# ndd -set /dev/fjgi 'adv_1000hdx_cap' 0

# ndd -set /dev/fjgi 'adv_100fdx_cap'

# ndd -set /dev/fjgi 'adv_autoneg_cap' 0

# ndd -set /dev/fjgi 'adv_autoneg_cap' 1

A list of parameter can be used with ndd command.

Table 4.2 A list of parameters

Parameter

Status

Meaning

Display parameter list

0 : Link down

Read only

link_status

1 : Link up

link_speed

Read only

1000 : 1000Mbps

100 : 100Mbps

10 : 10Mbps

link_mode

autonego

Read only

0 : Half Duplex communication

1 : Full Duplex communication

0 : Autonegotiation is Off.

1 : Auotnegotiation is On.

Parameter

Status

Meaning

0 : None (Not doing flow_control)

1 : LocSend (Can transmit pause

frame only)

flow_control

Read only

2 : Rem (Can receive pause frame

only)

3 : Sym (Can receive and transmit

pause frame)

cardtype

Read only

0 : SX(PW008GE4)

1 : T(PRIMEPOWER250/450

secondary LAN port,

PW008GE5/PW008QG1)

Instance Number

instance

Read and write

adv_10fdx_cap

Setting of 10Mbps/FullDuplex

0 : Invalid

1 : Valid (Default)

adv_10hdx_cap

adv_100fdx_cap

adv_100hdx_cap

adv_1000fdx_cap

adv_1000hdx_cap

adv_pauseTX

Read and write

Setting of 10Mbp/HalfDuplex

0 : Invalid

1 : Valid (Default)

Setting of 100Mbps/FullDuplex

0 : Invalid

1 : Valid (Default)

Setting of 100Mbps/HalfDuplex

0 : Invalid

1 : Valid (Default)

Setting of 1000Mbps/FullDuplex

0 : Invalid

1 : Valid (Default)

Setting of 1000Mbps/HalfDuplex

0 : Invalid

1 : Valid (Default)

Setting of transmit pause frame

0 : Invalid

1 : Valid (Default)

adv_pauseRX

Setting of receive pause frame

0 : Invalid

1 : Valid (Default)

adv_autoneg_cap

adv_role_cap

Setting of AutoNegotiation

0 : Invalid (Fixed mode)

1 : Valid (Default)

Setting of Role (Only 1000Mbps)

0 : Slave

1 : Master

2 : Auto (Default)

lp_10fdx_cap

Read only

Setting of link-partner

10Mbps/FullDuplex, when link up

with AutoNegotiation.

0 : Invalid

1 : Valid

Parameter

Status

Meaning

Setting of link-partner

10Mbps/HalfDuplex, when link up

with AutoNegotiation.

0 : Invalid

lp_10hdx_cap

Read only

1 : Valid

lp_100fdx_cap

lp_100hdx_cap

lp_1000fdx_cap

lp_1000hdx_cap

lp_pauseTX

Read only

Setting of link-partner

100Mbps/FullDuplex, when link up

with AutoNegotiation.

0 : Invalid

1 : Valid

Setting of link-partner

100Mbps/HalfDuplex, when link up

with AutoNegotiation.

0 : Invalid

1 : Valid

Setting of link-partner

1000Mbps/FullDuplex, when link up

with AutoNegotiation.

0 : Invalid

1 : Valid

Setting of link-partner

1000Mbps/HalfDuplex, when link up

with AutoNegotiation.

0 : Invalid

1 : Valid

Setting of link-partner transmit pause

frame, when link up with

AutoNegotiation.

0 : Invalid

1 : Valid

lp_pauseRX

Setting of link-partner receive pause

frame, when link up with

AutoNegotiation.

0 : Invalid

1 : Valid

lp_autoneg_cap

Setting of link-partner

AutoNegotiation, when link up with

AutoNegotiation.

0 : Invalid

1 : Valid

role_cap

jumbo

Read only

The present Role setting, when

1000Mbps.

0 : Slave

1 : Master

0 : JumboFrame function is not used.

1 : JumboFrame function is used.

Notes

Because of the specification of ndd utility, if two or more processes execute ndd command for a

fjgi interface at the same time, the value will be invalid.

If you execute ndd command repeatedly, fjgi interface performance may drop.

The values of link_speed, link_mode, autonego, flow_control are effective, only when link_status

= 1. If link_status = 0, these values are invalid.

The values of lp_10fdx_cap, lp_10hdx_cap, lp_100fdx_cap, lp_100hdx_cap, lp_1000fdx_cap,

lp_1000hdx_cap, lp_pauseTX, lp_pauseRX, lp_autoneg_cap are effective, only when

communication is done without Autonegation or Autonegation has not been established.

The values set by ndd command are invalid after reboot.

Since a setup of the operation mode by the ndd command becomes effective per physical interface,

when IPv4/IPv6 interface, SNA/FNA interface or the VLAN interface is being used, those are also

changed simultaneously.

4.3.2.4 Changing the setting of FCode

This section explains how to change the communication mode settings of FCode with examples.

If using PCI card (PW008GE4/PW008GE5/PW008QG1), please refer to the User’s guide attached with PCI

card.

It is needed when don't use autonegotiation for speed/duplex/flowcontrol or connecting partner cannot

support autonegotiation.

[FCode Force mode (Speed Fixed and disable autonegotiation) setting method]

Search for the device path of 'FJSV' or 'pwga' in the OK prompt screen.

Example

ok show-nets <Return>

a) /pci@83,2000/FJSV,pwga@1

b) /pci@1f,4000/network@1,1

q) NO SELECTION

Enter Selection, q to quit:

In this example, a) is the device path of the pwga (secondary LAN port)

With the following command, the device is specified.

ok cd /pci@83,2000/FJSV,pwga@1 <Return>

Implement the following commands, it will change from AutoNegotiation mode to Force mode.

This setting is invalid during OBP (Open Boot Prom). PW008GE4 (1000BASE-SX) can use only 3) speed

setting commands.

1) Role setting commands

slave-mode (slave)

2) Half Duplex setting commands

half-duplex (Half Duplex)

3) Speed setting commands

transfer-speed=100

transfer-speed=10

(Force Speed: 100Mbps)

(Force Speed: 10Mbps)

With the following command, the device setting information is displayed.

ok .prp <Return>

Example: PRIMEPOWER 450

ok cd /pci@83,2000/FJSV,pwga@1 <Return>

ok .prp <Return>

role

duplex

transfer-speed

cabinet-name

board-name

slave

half

0000000a

Cabinet#0

Motherboard#1

assigned-addresses 83002810 00000000 00100000 00000000 00010000

82002830 00000000 00110000 00000000 00010000

*1: Appears only when slave-mode is set.

*2: Appears only when half-duplex is set.

*3: Appears only when transfer-speed is set. [10Mbps:0xa, 100Mbps:0x64]

In the default, those *1,*2 and *3 are not appeared.

FCode Change back to default setting (AutoNegotiation mode) method

To return to the default setting after setting Fcode Force mode, do as follows:

On ok prompt screen, implement below command or power-off the system and power-on again. (Implement

only reset command cannot back to default setting.)

ok reset-all <Return>

4.3.2.5 Network Install

The procedure which uses fjgi interface to install the Solaris OS via a network from an installation server is

explained below.

Setup the install server and the clients to install the Solaris OS.

About setup of install server and clients, refer to "Solaris 10 Installation Guide."

In an installation server, execute the command shown below and check the root directory of a client

system. The root directory of a client system is described by the /etc/bootparams file of an

installation server.

# grep CLIENT_NAME /etc/bootparams <Return>

Example

# grep CLIENT_NAME /etc/bootparams

CLIENT_NAME root=SERVER_NAME:/netinstall/Solaris_10/Tools/Boot

install=

In the example, the root directory of this client is set to /netinstall.

CLIENT_NAME is the name of client system.

SERVER_NAME is the name of install server.

Information

When the root directory name is not described by the /etc/bootparams file, please refer to "Solaris 10

Installation Guide."

Check the ok prompt of OpenBootPROM is displayed on the client system.

If OS is booted, please stop a system by the shutdown(1M) command as follows.

# /usr/sbin/shutdown -i0 -g0 -y <Return>

Example

# /usr/sbin/shutdown -i0 -g0 –y

.......

(shutdown messages are displayed)

.......

Search the device path to the secondary LAN port with a show-nets command from ok prompt of a

client system.

ok show-nets <Return>

Example

ok show-nets

a) /pci@83,2000/FJSV,pwga@1

b) /pci@1f,4000/FJSV,pgsb@4

q) NO SELECTION

Enter Selection, q to quit:

From ok prompt of a client system, please start a client system using the secondary LAN port.

ok boot /pci@83,2000/FJSV,pwga@1 <Return>

Install the Solaris

4.3.2.6 Configuring VLAN interface

4.3.2.6.1 IEEE 802.1Q TagVLAN

The outline about a VLAN interface and directions required in order to setup VLAN interface are explained.

About IEEE802.1Q TagVLAN

By using VLAN, traffic management of two or more networks by a single physical interface becomes

possible. In order to create a VLAN environment, which resembles Figure 4.1 Example of VLAN

environment, a VLAN setup (VLAN ID and Tagged/Untagged) of a switch is required.

- No TagVLAN

PRIMEPOWER

- With TagVLAN

PRIMEPOWER

fjgi0

fjgi1

fjgi1000

fjgi2000

X.X.1.1

X.X.2.1

X.X.1.1

X.X.2.1

Secondary LAN NIC

secondary LAN

LAN Switch

To X.X.1.X

network

To X.X.2.X

network

To X.X.1.X

network

To X.X.2.X

network

(netmask: 255.255.255.0 )

Figure 4.1 Example of VLAN environment

By using VLAN, traffic management of two or more networks by the single physical interface can be done.

In order to create Figure 4.1 Example of VLAN environment, it is needed a VLAN setup (VLAN ID and

Tagged/Untagged) of a switch.

The vlan numbers fjgi interface can offer:

VLAN ID

1 – 4094 (are available)

1024 (can be created)

Max number of VLAN interfaces

4.3.2.6.2 How to create the VLAN interface

-To create the VLAN interface.

By setting the interface number to more than 1000, it becomes possible to distinguish VLAN interface

from physical interface.

How to set:

In case to create the "VLAN ID = 1" VLAN interface from instance = 0 physical interface.

Please refer to "4.3.1 Environment Setting ", and create the interface named as fjgi1000.

In case to create the "VLAN ID = 231" VLAN interface from instance = 3 physical interface.

Please refer to "4.3.1 Environment Setting", and create the interface named as fjgi231003.

Low 3 digits of VLAN interface number : physical instance number

Upper digits except low 3 digits of VLAN interface number: VLAN ID(1 -

4094).

VLAN interface number = VLAN ID * 1000 + physical instance number

Note:

z When connecting with a LAN switch, Tag VLAN needs to be set to a LAN switch. (Please refer to a

LAN switch specification which supported Tag VLAN.)

z Please use a VLAN interface only with TCP/IP protocol. Please do not use it with SNA/FNA and OSI

protocol.

z Since PRIMECLUSTER GLS 4.1A20 or previous version do not correspond to a VLAN interface,

please do not use VLAN interface with those. PRIMECLUSTER GLS 4.1A30 or later (only Fast

switching mode or NIC switching mode) can use VLAN interface.

z VLAN function can be used only in Solaris 8, Solaris 9 or Solaris 10.

z In the environment where PRIMECLUSTER has been installed, in case you use a VLAN interface,

please check that the "SMAWdtcp" package has not been installed. When installed, please remove

that package or do not use a VLAN interface.

z A VLAN interface uses about 700Kbyte at MTU size 1514byte or 900Kbyte at MTU size 9014byte of

memory (at the time of no communicating). Therefore, when creating two or more VLAN interfaces

and performing communication, a slowdown may occur due to the shortage of resources depending on

the system configuration.

4.4 Messages

Messages by fjgi driver are listed in the following table.

Table 4.3 Console Messages of the fjgi driver

Message

Cause

Workaround

network connection down

Link down was happened.

When Link down not intended is

generated, abnormality of the

transmission line is thought. Please

confirm the specification of the cable

and the switch.

network connection up

using portA

Connected

There is no problem.

-speed

speed:

1000: Transmitting and receiving by

1000Mbps

autonegotiation:

duplex mode:

flowctrl:

100: Transmitting and receiving by

100Mbps

( role:

)*1

10: Transmitting and receiving by

10Mbps

-autonegotiation

on: Set in autonegotiation

off: Not setting in autonegotiation

-duplex mode

full: Transmitting and receiving by

full duplex

half: Transmitting and receiving by

half duplex

-flowctrl

symmetric: Doing Flowcontrol each

other

remote send: Accept Flowcontol only

from

the

other.

local send: Doing only own

Flowcontrol

none: Not doing Flowcontrol.

Message

Cause

Workaround

-role

master: Communicate as Master.

slave: Communicate as Slave.

Error found in the parameter of

definition file (fjgi.conf).

Illegal value for

[ParameterName].

fail to FragCount

Check whether the definition file is

correct.

Abnormalities were found in message

fragmentation.

Please try re-activation of interface. If

this is not a cause, report that to your

authorized dealer.

Multiple fragments in

descriptor area

Abnormalities were found in message

fragmentation.

bind to 802.2 failed

pci_config_setup_failed

Using protocol was different.

PCI configuration space couldn't be

obtained.

Installed physical memory may be

insufficient or hardware is failed. If

memory insufficient is not a cause,

report that to your authorized dealer.

Installed physical memory may be

insufficient. If this is not a cause,

report them to your authorized dealer.

soft state alloc failed.

Driver internal resources cannot be

obtained.

Allocation of descriptor

memory failed

Driver internal resources on DMA

couldn't be allocated.

fail to

DMA resources couldn't be allocated.

ddi_dma_bind_handle

Could not allocate DMA

handle

DMA resources couldn't be allocated.

Could not bind descriptor

memory

fjgiallocthings failed !!

Unable to map adapter

registers.

DMA resources couldn't be allocated.

Mapping of register went wrong.

ddi_get_iblock_cookie-E

RROR

Interrupt resources couldn't be

allocated.

Unable to install interrupt

routine.

Unable to install interrupt routine.

Interrupt resources couldn't be

allocated.

ddi_create_minor_node

failed

Creation of a device file went wrong.

ATTACH failed

Driver resources couldn't be allocated.

OPEN failed: ENOMEM

CLOSE: stream==NULL

fjgi_param_register error

Driver resources couldn't be freed.

The ndd parameter couldn't be

allocated.

fjgi_board_alloc_mem

failed !!

Driver resources couldn't be allocated.

Error filling TX ring

Transmit resources couldn't be

allocated.

status block can't alloced !!

Status information resources couldn't

be allocated.

status block can't alloced

handle!

Status information resources couldn't

be allocated.

status block can't bind

handle!

Status information resources couldn't

be allocated.

Message

fjgi_board_init_mem:

Failed to dma_alloc !!

statistick block can't

alloced!!

Cause

Workaround

DMA resources cannot be allocated.

Statistic

information

resources

hwmib block can't alloced

handle!

couldn't be allocated.

hwmib block can't bind

handle!

fjgi_board_init_mem:

Failed to dma alloc2!!

Illegal value for

LinkSpeed_A

DMA resources cannot be allocated.

The value of LinkSpeed_A of

setting file was wrong.

Please set a right value of

LinkSpeed_A.

fjgi_hw_init:Step 22

failed !!

Hardware failure was occurred. Please

report to your authorized dealer.

fjgi_hw_init:Step 23

failed !!

The error occurred in initialization.

fjgi_hw_init:Step 43

failed !!

Illegal combination of

values AutoNeg and Speed

Except for the above

Auto speed couldn't be set in fixed

mode.

Please set value except Auto in

LinkSpeed_A.

There is the possibility that is Driver

internal bug or Hardware failer.

Report that to your authorized dealer.

*1: This appears only when 1000Mbps is used.

Table 4.4 Machine management report messages

Message

Meaning

The error occurred in initialization of

an adapter.

Workaround

There is the possibility of hard

abnormalities. Report them to your

authorized dealer.

fjgi_device_check:

fjgi_hw_deinit failed !

fjgi_device_check:

fjgi_hw_init failed !

pci_config_setup_failed

Acquisition of PCI configuration

space went wrong.

There is the possibility of the shortage

of resources or hard abnormalities.

Please check the empty space of a

memory. Report them to your

authorized dealer.

4.5 Using GigabitEthernet on a Cluster System

4.5.1 Support for Cluster System

The contents of support at the time of using by the cluster system, the environmental setting method, etc. are

explained.

GigabitEthernet has the following functions on the cluster system:

z Operation standby mode

z 1:n operation standby mode

z Mutual standby mode

z TCP/IP communication

z Using to the cluster interconnect (PRIMECLUSTER). (But, using in a JumboFrame or as

VLAN interface can not be performed)

The support ranges of takeover form are the following ranges.

PRIMECLUSTER

Taking over form

Support

MAC

Not Supported

(Supported V4.1A10 or later)

Not Supported

IP+ MAC

(Supported V4.1A10 or later)

Caution

Using over operation management LAN (LAN used for cluster operation management view) cannot be

supported.

4.5.2 Cluster Environment Setup Procedure

Set up the cluster environment in the same way as for the other LAN driver, follow "Setting up the

Network" in the "PRIMECLUSTER manual". Note that the secondary LAN port (fjgi interface) resource

name is "fjgi".

4.5.3 Notes in Cluster Environment

Please use "PRIMECLUSTER 4.1 or later" for a cluster system to use fjgi interface. If an earlier version of

cluster system is used, fjgi interface cannot be used for private LAN nor for IP address fail-over on public

LAN.

When using fjgi interface as the cluster interconnect, please perform the following command before initial

composition setup.

# ifconfig fjgi* plumb <Return>

* is instance-number

Then, please execute the following command and check that a MTU size of fjgi interface is not 9000byte

(JumboFrame is not set up).

# ifconfig fjgi* <Return>

* is instance-number

GigabitEthernet in a JumboFrame function cannot be used to the cluster interconnect.

GigabitEthernet as VLAN interface cannot be used to the cluster interconnect.

Chapter 5 Precautions

This chapter explains necessary precautions.

5.1 Limitations and known bugs for Solaris

This chapter explains limitations and known bugs for Solaris.

Table 5.1 Limitations and known bugs for Solaris

Limitations and known bugs

Recommended actions

Remarks

The following patches are required to be installed in the following order after installation Solaris 10

of Solaris 10. In particular 119254-06 must be installed before 119254-19 or later or

equivalent. If the installation order is not followed problems may occur.

3/05 and

Solaris 10

1/06 only

E.g. non-global zone cannot be installed, optional packages and patches cannot be

installed correctly.

1. 120719-01 or later SunOS 5.10: gzip patch

2. 121296-01 or later SunOS 5.10: fgrep Patch

3. 118815-03 or later SunOS 5.10: awk patch

4. 120900-03 or later SunOS 5.10: libzonecfg Patch

5. 121133-02 or later SunOS 5.10: zones library and zones utility patch

6. 119254-06

SunOS 5.10: Install and Patch Utilities Patch

7. 119254-19 or later: SunOS 5.10: Install and Patch Utilities Patch

8. 119317-01 or later: SunOS 5.10: SVr4 Packaging Commands (usr) Patch

9. 119042-02 or later: SunOS 5.10: patch usr/sbin/svccfg

Install these patches as soon as possible after installation of Solaris 10. Also make sure

of the following notes when applying/removing any patches.

Read README files of the patches before applying them.

Remove patches in the global zone and not in the non-global zone if needing to

remove patches from the system.

The following patch is required to use the NFS version 4.

NFS version 4 has limitations.

Solaris 10

3/05 only

119828-04 or later SunOS 5.10: nfssrv Patch

Install the above patches.

For Dynamic Reconfiguration (DR) and PCI Hot Plug, the following patches are

required to be installed.

Solaris 10

3/05 only

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

118822-08 or later SunOS 5.10: kernel Patch

913732-01 or later Enhanced Support Facility 2.5: Dynamic configuration core

command patch

Install the above patches.

For CPC (CPU Performance Counter) (cpustat(1M), cputrack(1M), libcpc(3LIB)), the

following patches are required to be installed.

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

118822-08 or later SunOS 5.10: kernel Patch

Install the above patches.

Solaris 10

3/05 only

Limitations and known bugs

Remarks

Recommended actions

For growfs(1M), the following patch is required to be installed.

120254-02 or later SunOS 5.10: ufs Patch

Install the above patches.

Solaris 10

3/05 and

1/06 only

Solaris 10

3/05 only

When using any of the following PGX cards and monitors, the display monitor does not

enter standby mode if you haven’t logged on from CDE (Common Desktop

Environment).

PGX frame buffer (X3660A)

PGX64 frame buffer (X3768A/A-A)

Second monitor (X7128A/AA/AJ, X7128B)

To use PGX32 frame buffer (X3668A/A-A, X3668B/B-A/B-B), the following patches

are required to be installed.

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

119309-03 or later SunOS 5.10: PGX32 Graphics

The above removes the PGX frame buffer and monitor limitation. When using the PGX,

PGX64 frame buffer and a second monitor, ensure you turn off the power to the display

monitor when the monitor is not in use.

Some optional drivers are not yet supported with Solaris 10.

Contact your support engineer for details.

/etc/resolv.conf file might not be created when you select “DNS” name service during

OS installation.

Solaris 10

3/05 only

Install the following patch.

119252-02 or later SunOS 5.10: System Administration Applications Patch

svccfg(1M) might be terminated when editprop subcommand is used.

# svccfg

svc:> select /network/inetd

svc:/network/inetd> editprop

svccfg: unrecognized character "

Use setprop subcommand of svccfg(1M) instead of editprop to modify properties of

service.

For Kernel debugger (kmdb(1), mdb(1) -K) on SPARC64 GP, the following patches are

required to be installed.

Solaris 10

3/05 only

118822-23 or later SunOS 5.10: kernel Patch

Install the above patch.

Solaris Live Upgrade using lu(1M) and lucreate(1M) might fail with the following error

when creating a mirror disk.

# lucreate -m /:/dev/md/dsk/d10:ufs,mirror –m \

/:/dev/dsk/c1t2d0s0,d11:attach -m/:/dev/dsk/c1t3d0s0,d12:attach -n newBE

Discovering physical storage devices

Discovering logical storage devices

Cross referencing storage devices with boot environment configurations

Determining types of file systems supported

Validating file system requests

ERROR: file system option <mirror> is not valid on this system

ERROR: cannot use file system configuration </:/dev/md/dsk/d10:ufs,mirror>

ERROR: cannot create new boot environment using file systems as configured

ERROR: please review all file system configuration options

ERROR: cannot create new boot environment using options provided

Limitations and known bugs

Remarks

Recommended actions

Solaris Live Upgrade is not supported with RAID-1(mirror)

Custom Jumpstart installation for mirror volumes using a profile that has filesys keyword

of both original and target disk, will fail and the following error message will be shown

at first boot after the installation.

metattach: host-name: dx: can't attach labeled submirror to an unlabeled mirror

Create a one-way mirror in the profile file. Define a primary submirror Disk as filesys

keyword. Then attach additional submirrors with the finish script to create a two-way

mirror.

Do not include cylinder 0 of the disks in meta devices.

The disks which constitute a two-way mirror should have the same specification

and configuration. Examples of a profile file and a finish script are shown as

follows:

1) profile file

# profile keywords

profile values

# ----------------

-----------------

filesys

c0t1d0s4 0:1 unnamed

filesys

mirror:d100 c0t1d0s0 200 /

c0t1d0s3 size 8192 count 2

metadb

2) finish script

#!/bin/sh

prtvtoc /dev/rdsk/c0t1d0s2 > /a/c0t1d0.vtoc

fmthard -s /a/c0t1d0.vtoc /dev/rdsk/c0t0d0s2

metadb -a -c 2 c0t0d0s3

metainit d102 1 1 c0t0d0s0

metattach d100 d102

in.routed might generate a core dump under the following conditions.

1) No entry of localhost name in /etc/hosts and

2) No /etc/defaultrouter file and

3) Reboot system

Install the following patch.

119712-03 or later SunOS 5.10: usr/sbin/in.routed Patch

Package remove using pkgrm(1M) might fail if the package provides SMF services and

the package has been removed and installed repeatedly.

## Removing pathnames in class <manifest>

svc:/system/fjsvmadm-evhandsd:default remains enabled; aborting

pkgrm: ERROR: class action script did not complete successfully

Limitations and known bugs

Remarks

Recommended actions

Follow these instructions to remove a package. (For example SUNWabc)

1) Re-install the package to system.

# pkgadd SUNWabc

2) Find out path name of SMF service manifest in the package.

# grep SUNWabc /var/sadm/install/contents | \

grep /var/svc/manifest | awk '{ print $1 }' | grep xml

/var/svc/manifest/system/abc.xml

3) Run the following command.

# svccfg -s smf/manifest delpg {$prp_mfst}

$prp_mfst strings must be specified by removing first “/” srting of the manifest path

name with all “/” and ”.” replaced by “_”. Here is an example for

/var/svc/manifest/system/abc.xml at step 2).

4) Remove package.

# pkgrm SUNWabc

"netstat -i" results maybe corrupted when the network interface on the system is

unplumb/plumb by ifconfig(1M), if the system has multiple physical network interfaces.

This problem is solved by the following patch.

Solaris 10

3/05 only

119990-01 or later SunOS 5.10: netstat patch

Network instance number on Quad Fast Ethernet and Dual Gigabit Ethernet Card might

be assigned randomly to physical network interfaces when the card is installed after OS

installation.

Solaris 10

3/05 only

# grep fjqe /etc/path_to_inst

"/pci@81,4000/pci@2/FJSV,pl4b@4" 0 "fjqe"

"/pci@81,4000/pci@2/FJSV,pl4b@5" 3 "fjqe"

"/pci@81,4000/pci@2/FJSV,pl4b@6" 5 "fjqe"

"/pci@81,4000/pci@2/FJSV,pl4b@7" 7 "fjqe"

"/pci@85,4000/pci@2/ethernet@4" 1 "fjqe"

"/pci@85,4000/pci@2/ethernet@5" 2 "fjqe"

"/pci@85,4000/pci@2/ethernet@6" 4 "fjqe"

"/pci@85,4000/pci@2/ethernet@7" 6 "fjqe"

Install the following patch before installation of Quad Ethernet Cards or Quad

GigabitEthernet Cards.

118822-23 or later SunOS 5.10: kernel Patch

mpstat(1M) takes an extra interval time at the ending of the command when the

command is executed with interval and count.

Solaris 10

3/05 only

This problem is solved by the following patch.

118373-01 or later SunOS 5.10: mpstat Patch

pkgadd(1M) -R command might fail when target directory is on a remote NFS server.

This problem is solved by the following patch.

Solaris 10

3/05 only

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

The following message might be displayed if during optional driver package installation

the PCI Card is not installed.

chgrp: //dev/{driver_name}: No such file or directory

There is no impact on the system. The driver and required PCI Card work after a system

reboot.

Limitations and known bugs

Remarks

Recommended actions

"Internet (IP) Address” in “Workstation information” of CDE environment is shown

with a IPv6 unicast address as “::FFFF:***.***.***.***”.

This problem is solved by the following patch.

Solaris 10

3/05 only

119284-01 or later CDE 1.6: sdtwsinfo patch

During system shutdown, some messages similar to the following might be displayed on

the console.

syslogd: /var/opt/FJSVmadm/evh/evh_pipe: Broken pipe

NFS fsinfo failed for server xxxx: error 5 (RPC: Timed out)

Ignore the message. It has no impact on the system.

ifconfig(1M) displays “Need net number for mask” when the command is executed with

plumb and up option at the same time.

# ifconfig hme0 plumb host netmask + broadcast + up

Need net number for mask

Setting netmask of hme0 to 255.255.255.0

Ignore the “Need net number for mask” message. The message will not be displayed

when plumb and up option are specified separately.

# ifconfig hme0 plumb

# ifconfig hme0 host netmask + broadcast + up

Setting netmask of hme0 to 255.255.255.0

Solaris Management Console(SMC) 2.1, User Performance (“System Status”->

“Performance” -> “System” -> “Users”) does not show any information and the

following error messages displays on the console.

CIM_ERR_FAILED:

java.lang.UnsupportedClassVersionError:

com/sun/wbem/solarisprovider/srm/Solaris_ActiveUser (Unsupported major.minor

version 49.0)

Use ps(1), prstat(1M) and pgrep(1) to check and report active process statistics instead of

the SMC feature.

The following message might be shown during Solaris 10 installation.

application/font/fc-cache:default failed (see 'svcs -x' for details)

Ignore the message. It has no impact on the system.

The following message might be shown during Solaris 10 installation.

/usr/sbin/pmconfig: cannot open "/dev/pm": No such file or directory

Ignore the message. It has no impact on the system.

After Solaris 10 installation is finished, the following messages might be shown at first

boot.

java[16552]: [ID 874347 user.error] libpkcs11: open /var/run/kcfd_door: No such

file or directory

java[16552]: [ID 523743 user.error] libpkcs11:

/usr/lib/security/pkcs11_softtoken.so unexpected failure in ELF signature

verification. System may have been tampered with. Cannot continue parsing

/etc/crypto/pkcs11.conf

Ignore these messages. They have no impact on the system.

Limitations and known bugs

Remarks

Recommended actions

There are some differences in behavior of MIB between SNMP agents of SMA and

SEA.

sysObjectID (MIB to show system device type)

SEA shows the value as “enterprises.42.2.1.1”, however, SMA shows the value as

“netSnmpAgentOIDs.3”.

snmpEnableAuthenTraps (MIB to show permission of trap of authorization

failure)

SEA shows the value as “enabled(1)”, however, SMA shows the value as

“disabled(2)”.

ifSpeed (MIB for supported bandwidth speed of the network interface)

The results may be different between SEA and SMA.

ifType

The results may be different between SEA and SMA.

Use SEA instead of SMA. The following disables SMA and enables only SEA.

1) Stop SMA and rename rc startup script of SMA.

# /etc/init.d/init.sma stop

# mv /etc/rc3.d/S82initsma /etc/rc3.d/_S82initsma

2) Rename /etc/rc3.d/S75seaport

# mv /etc/rc3.d/S75seaport /etc/rc3.d/_S75seaport

3) Set port number from 16161 to 161 for SEA in /etc/snmp/conf/snmpdx.reg

4) Create configuration files for SEA

# cp /etc/snmp/conf/mibiisa.rsrc- /etc/snmp/conf/mibiisa.rsrc

# cp /etc/snmp/conf/snmpdx.acl /etc/snmp/conf/mibiisa.acl

5) Restart SEA

# /etc/init.d/init.snmpdx stop

# /etc/init.d/init.snmpdx start

The following error messages might be logged in SMF service log file.

ERROR: Restarter: Ignoring REMOVE_INSTANCE command for unknown

service svc:/TEMP/{FMRI}.

Ignore these messages. They have no impact on the system.

Packages that contain SMF service can not be removed using pkgrm(1M)

-R option from the other boot system,and the following messages are displayed during

the next system boot.

SunOS Release 5.10 Version Generic 64-bit

Use is subject to license terms.

Hostname: host-name

Starting SCF daemon.

svc.startd[7]: svc:/system/fjsvabc:default: Method "/lib/svc/method/fjsvabc" failed

with exit status 1.

svc.startd[7]: svc:/system/fjsvabc:default: Method "/lib/svc/method/fjsvabc" failed

with exit status 1.

svc.startd[7]: svc:/system/fjsvabc:default: Method "/lib/svc/method/fjsvabc" failed

with exit status 1.

[ system/fjsvabc:default failed (see 'svcs -x' for details) ]

-R option for pkgrm (1M) has a limitation for packages that include SMF service.

Remove the package on the local file system after booting the remote image.

Limitations and known bugs

Remarks

Recommended actions

A /noautoshutdown file might be created at first boot after Solaris 10 installation. If

/noautoshutdown file exists power management and power saving could be disabled.

Check for and remove /noautoshutdown file if you need to enable power management

and power saving features.

Patch installation may fail with the following error message when the patch binary is

located in /(root) directory.

Solaris 10

3/05 only

# patchadd xxxxxx-xx

Assertion failed: *a_gzPath != ‘\0’, file zones.c, line 2113

This problem is solved by the following patch.

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

Installation of non-global zone will fail if the patches that have .us and .u instances,

including the following patches, are installed in the system.

118822-02 SunOS 5.10: Kernel Patch

Solaris 10

3/05 only

119309-03 SunOS 5.10: PGX32 Graphics

118708-03 SunOS 5.10: Sun XVR-1200 and Sun XVR-600 Graphics Accelerator

Patch

118708-04 SunOS 5.10: Sun XVR-1200 and Sun XVR-600 Graphics Accelerator

Patch

If this problem occurs, the following errors show up when installing non-global zones

and the operation fails.

# zoneadm -z zone01 install

Preparing to install zone <zone01>.

Creating list of files to copy from the global zone.

Copying <129399> files to the zone.

Initializing zone product registry.

Determining zone package initialization order.

Preparing to initialize <1119> packages on the zone.

Initialized <1119> packages on zone.

Zone <zone01> is initialized.

Installation of these packages generated errors: <SUNWcakr>

Immediately install the following patch as soon as Solaris 10 is installed.

119254-06 or later SunOS 5.10: Install and Patch Utilities Patch

If you have already installed 118822-02, 119309-03, 118708-03, 118708-04 without

119254-06, follow the below procedure to recover the system.

1) Install 119254-06 (or later).

2) Install 118822-08 (or later) and 118708-06 (or later).

3) If 119309-03 has been installed, remove 119309-03 and re-install 119309-03.

4) Reboot system.

During system booting, port monitor (like ttymon) log files are deleted and the following Solaris 10

error message is displayed. Moreover, once this error message is output, no further log is 3/05 only

saved.

ttymon cannot create log file "/var/saf/zsmon/log": No such file or directory

This problem is solved by the following patch.

118868-01 or later SunOS 5.10: ttymon patch

Limitations and known bugs

Remarks

Recommended actions

On PRIMEPOWER 200 (360MHz or higher) / 400(360MHz or higher)/600/650(except

SPARC64 V)/800/850(except SPARC64 V)/1000/1500(except SPARC64 V)/2000, both

syslog message and FMA message are displayed when a Memory 1 bit error(CE) occurs.

e.g.)

FJSV,SPARC64-IV: [ID 635115 kern.notice] Softerrors:

FJSV,SPARC64-IV: [ID 663638 kern.notice] Seen 0 Intermittent and 1 Corrected

Softerrors

FJSV,SPARC64-IV: [ID 909492 kern.notice] from SIMM C0S00-SLOT#B00 (synd =

0x13, addr = 0x75268280).

FJSV,SPARC64-IV: [ID 805745 kern.notice] Softerror: Persistent ECC Memory Error

FJSV,SPARC64-IV: [ID 306888 kern.notice] Corrected SIMM C0S00-SLOT#B00

FJSV,SPARC64-IV: [ID 518845 kern.notice] Detected by Module ID = 0x87 (135)

unix: [ID 618185 kern.notice] NOTICE: Scheduling removal of page

0x00000000.75268000 fmd: [ID 441519 daemon.error]

SUNW-MSG-ID: SUN4US-8000-7D, TYPE: Fault, VER: 1, SEVERITY: Minor

EVENT-TIME: Mon Feb 21 21:26:51 JST 2005

PLATFORM: FJSV,GPUSC-M, CSN: -, HOSTNAME: colm2

SOURCE: cpumem-diagnosis, REV: 1.3

EVENT-ID: aaf47b25-cfd2-4984-f598-e81bcbba274f

DESC: The number of single bit errors associated with this memory module continues

to exceed acceptable levels. Refer to

http://www.fujitsu.com/global/services/computing/server/unix/prmpwr_msg/SUN4US

-8000-7D for more information.

AUTO-RESPONSE: An attempt will be made to remove this memory page from

service.

IMPACT: Total system memory capacity will be reduced as pages are retired.

REC-ACTION: Schedule a repair procedure to replace the affected memory module.

Use fmdump -v -u <EVENT_ID> to identify the module.

Ignore these syslog messages. They have no impact on the system.

System does not display CPU error messages via FMA if the CPU at fault has not been

replaced; even if the CPU detected the same error previously. This problem only happens

on systems installed with SPARC64 V.

This is just a messaging issue. It has no impact on the system as such.

After an upgrade to install Solaris 10, the following message is displayed when format

(1M) command is executed.

# format

Unknown controller 'MD21' - /etc/format.dat (xx)

Searching for disks...done

There is no impact on the system. The message is not displayed when the following

entries are removed from /etc/format.dat.

disk_type = "XXXX" \

: ctlr = MD21 \

: ncyl = XXXX : acyl = X : pcyl = XXXX : nhead = X : nsect = XX \

: rpm = XXXX : bpt = XXXXX

Limitations and known bugs

Remarks

Recommended actions

The -R option for pkgadd(1M), pkgrm(1M), patchadd (1M), and patchrm(1M) utilities

have limitations when zones are included in the remote target image.

When a package and/or patch needs to be removed from non-global zone, install the

following patch before patch/package removal.

Solaris 10

3/05 only

119254-19 or later SunOS 5.10: Install and Patch Utilities Patch

LiveUpgrade installation and Solaris Upgrade Installation of non-global zones have

limitations.

Announcement will be made when support for this feature is enabled.

When you install zone using zoneadm(1M), the following error messages with FJSVdr

or FJSVparam is displayed.

# zoneadm -z myzone install

Preparing to install zone <myzone>.

Creating list of files to copy from the global zone.

Copying <2784> files to the zone.

Initializing zone product registry.

Determining zone package initialization order.

Preparing to initialize <1081> packages on the zone.

Initialized <1081> packages on zone.

Zone <myzone> is initialized.

Installation of these packages generated errors: <FJSVdr>

Installation of <19> packages was skipped.

The file </work/myzone/root/var/sadm/system/logs/install_log> contains a log of the

zone installation.

They have no impact on the system.

During boot system, the following warning message may be displayed.

WARNING: Time of Day clock error: reason [Reversed by 0x2]. -- Stopped

tracking Time Of Day clock.

Ignore the message. It has no impact on the system.

The following option for patchadd(1M) does not work.

-d , -u, -n, -B, -k, -P, -x

There is no workaround now.

The following message might be displayed on the console.

snmpXdmid: [ID 216524 daemon.error] Registration with DMI failed. err = 831.

If there is no map file except test.map in /var/dmi/map, it means DMI is not used, then

disable DMI service provider in the following way.

On Solaris 10 OS 6/06 and earlier.

# mv /etc/rc3.d/S77dmi /etc/rc3.d/_S77dmi

On Solaris 10 11/06 and later.

# svcadm disable svc:/application/management/dmi:default

When removing a patch with patchrm(1M), you may fail to remove it with the following

messages.

First RE may not be null

Patchrm is terminating.

WARNING: patchrm returned <21>

Install the following patch.

119254-19 or later SunOS 5.10: Install and Patch Utilities Patch

Limitations and known bugs

Remarks

Recommended actions

I Installation and removal of a package or patch may fail in a system with non-global

zones, under the following conditions.

non-global zone is configured using zonecfg(1M), but not installed by

zoneadm(1M).

non-global zone is just booted by zoneadm(1M) after installation, without any

non-global zone settings such as host name and root password.

non-global zone is being installed or removed by zoneadm(1M).

Make sure the initial setup like hostname and password configuration of all non-global

zones has been done and all the non-global zones are in “running” state before

installation or removal of package and patch

The following message may be displayed during an upgrade installation from Solaris 9.

network/nfs/server:default failed repeatedly (see 'svcs -x' for details)

Ignore the message. It has no impact on the system.

5.2 Specification and incompatibility for Solaris

This chapter explains specification and incompatibility for Solaris.

Table 5.2 Specification and incompatibility for Solaris

Specification and incompatibility

Remarks

Recommended actions

Solaris 10 provides Solaris Service Manager (SMF) to enhance the UNIX startup script

(so called RC script) and system daemons during boot up. Due to this new feature, there

is some unexpected and incompatible behavior compared to Solaris 9 or older versions.

Refer to SMF documentation as follows and smf(5) manual page.

Solaris 10 System Administrator Collection

>> System Administration Guide: Basic Administration

http://docs.sun.com/app/docs/doc/817-1985/

/usr/bin/ln has been changed in Solaris 10.

Solaris 9 or older versions.

/usr/bin/ln will overwrite file if target file already exists.

Solaris 10.

/usr/bin/ln will fail if target file already exists.

The same behavior occurs with Solaris 9 or older versions if the -f option is used with

the ln command.

If the DHCP sever has multiple network interfaces, it may not be able to deliver an IP

address to a DHCP client because of DHCP server settings.

pntadm(1M) -s option needs to be specified with "uname -n" or the IP address of the

primary network interface. If a non-primary IP address is specified with pntadm -s,

DHCP server will not be able to deliver an IP address to the DHCP client.

Specification and incompatibility

Remarks

Recommended actions

nslookup(1M) results on Solaris 10 differ from those of Solaris 9 or older versions.

e.g.)

Solaris 9

# nslookup

> aaa213.sample.test.com.

Server: aaat11.sample.test.com

Address: xxx.xxx.xxx.nnn

Name: aaa13.sample.test.com

Address: xxx.xxx.xxx.mmm

Aliases: aaa213.sample.test.com, aaa113.sample.test.com

Solaris 10

# nslookup

> aaa213.sample.test.com.

Server: xxx.xxx.xxx.nnn

Address: xxx.xxx.xxx.nnn#xx

aaa213.sample.test.com canonical name = aaa113.sample.test.com.

aaa113.sample.test.com canonical name = aaa13.sample.test.com.

Name: aaa13.sample.test.com

Address: xxx.xxx.xxx.mmm

* aaa113.sample.test.com is an alias of aaa13.sample.test.com

* aaa213.sample.test.com is an alias of aaa113.sample.test.com.

1) Physical device (/devices) entries for hard disk drive are removed at system

reboot if the hard disk driver is removed from the system before system reboot.

e.g.)

# cd /dev/rdsk

# ls -l c1t6d0*

lrwxrwxrwx 1 root root 48 May 28 11:19 c1t6d0s0 ->../../devices/pci@1f,0

/pci@5/scsi@2/sd@6,0:a,raw

# ls -lL c1t6d0*

c1t6d0s0: No such file or directory

2) “iostat –E” and “drcstat –device” may not show all connected physical devices

right after system boot.

This behavior of devfs on Solaris 10, is to be expected. /devices entries are updated

dynamically when each device is actually attached. As a result, device entries that have

not been attached after system boot, are not displayed by those commands.

If non-global zones are not booted, patchadd(1M) and patchrm(1M) will make all

non-global zones boot up temporarily during patch installation and removal. As a result

the total time for patch installation and removal will take much longer than booting up in

single user mode.

Boot all non-global zones in single user mode before installing and removing patches.

Example)

global# zoneadm –z myzone boot -s

Specification and incompatibility

Remarks

Recommended actions

The following messages may be displayed during patch installation using patchadd(1M).

Package {PKGNAME} from directory {PKGNAME} in patch {PATCHID} is not

installed on the system.

Architecture for package {PKGNAME} from directory {PKGNAME} in patch

{PATCHID} differs from the package installed on the system.

Example)

# patchadd 118822-18

....

Validating patches...

Loading patches installed on the system...

Done!

Loading patches requested to install.

Package SUNWcti2 from directory SUNWcti2.u in patch 118822-18 is not installed

on the system.

Package SUNWdrcr from directory SUNWdrcr.u in patch 118822-18 is not

installed on the system.

Package SUNWidn from directory SUNWidn.u in patch 118822-18 is not installed

on the system.

Architecture for package SUNWdrr from directory SUNWdrr.u in patch 118822-18

differs from the package installed on the system.

Package SUNWwrsd from directory SUNWwrsd.u in patch 118822-18 is not

installed on the system.

Architecture for package SUNWefc from directory SUNWefc.u in patch 118822-18

differs from the package installed on the system.

Package SUNWwrsm from directory SUNWwrsm.u in patch 118822-18 is not

installed on the system.

Architecture for package SUNWcakr from directory SUNWcakr.u in patch

118822-18 differs from the package installed on the system.

Architecture for package SUNWcpr from directory SUNWcpr.u in patch 118822-18

differs from the package installed on the system.

Done!

.....

Ignore these messages. They have no impact on the system.

Specification and incompatibility

Remarks

Recommended actions

In Solaris10, rsh(1) will output messages according to the value of the LANG variable

that is defined in /etc/default/init. However, in the previous Solaris versions, the system

will use the LANG=C locale for a command, that is run at the server side by rsh(1),

without mentioning any specific value for the LANG variable.

In Solaris 10, if the LANG variable is not set explicitly to a process that is executed by

inetd(1M), the process will succeed the value of this variable from inetd(1M). To stop a

process from succeeding the value of the LANG variable from inetd(M), change the

setup with the following command:

e.g.) in case of rsh(1)

#inetadm -m network/shell:default inherit_env=FALSE

For details of the setting, see the inetd(1M) man page.

Patch installation using patchadd(1M) will fail if the system has any zones and if the

patch archive is on NFS server.

Copy the patch archive from NFS server to local file system such as /tmp and install the

patch.

Esoteric CD Player X 01 D2 User Manual
Faber Ventilation Hood Inca Pro 30 User Manual
Fellowes Paper Shredder 49C User Manual
Fire Magic Gas Grill L C2 01805 User Manual
Furuno Telephone FS 1502 User Manual
Gas Fired Products Electric Heater PTS User Manual
Generac Portable Generator 0043736 0046265 User Manual
GE Telephone 29870 Series User Manual
GE Vacuum Cleaner 106679 User Manual
Goodman Mfg Air Conditioner SX14 R 410A User Manual