[Return to Library]  [TOC]  [PREV]  SECT--  [NEXT]  [INDEX] [Help]

C    Installation Enhancements

This appendix describes enhancements to the Digital UNIX Version 4.0B full (default or custom) installation process and to the installation cloning process. Table C-1 summarizes the installation enhancements in this release.
Table C-1: Summary of Digital UNIX Version 4.0B Installation Enhancements
EnhancementApplies To:
The installation process searches for and invokes user-supplied files to enable customizations on the system to be installed. The files can be on diskette, a RIS server, the /var/tmp directory on your system, or on CD-ROM.  Full Installations and Installation Cloning 
Administrators can modify the configuration description file (CDF) to achieve an unattended installation cloning process  Installation Cloning 

The following information is included in this appendix:


[Return to Library]  [TOC]  [PREV]  SECT--  [NEXT]  [INDEX] [Help]

C.1    Installation Cloning Overview

Installation cloning allows an administrator to replicate the installation configuration from a model system that is already installed with Digital UNIX Version 4.0B onto one or more systems with the same or similar hardware configurations.

When a system is installed with Digital UNIX Version 4.0B, a configuration description file (CDF) is generated that contains the results of the questions answered during the installation. This file is located on the installed system in the /var/adm/smlogs directory under the file name install.cdf. The CDF contains all the configuration information required to perform an initial system installation on a client system.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.1.1    Prerequisites for Installation Cloning

The only prerequisite for installation cloning is that the system to be installed by the installation cloning process has the same disk configuration as the system where the CDF was generated. This means that the disks used for the / (root), usr, and var file systems and swap areas on both systems must have the same disk type and the same device name.

It is possible, however, to support slight differences in configuration. Section C.7.1 describes these acceptable differences.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.1.2    Benefits of Installation Cloning

The benefits to using installation cloning to mass-install systems are:


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.1.3    Installation Cloning Features

The files necessary for the installation cloning process can be placed on a diskette, the /var/adm/ris/clients/sets/profile_set directory on a RIS server or in the /isl directory on a CD-ROM or extracted RIS area. A CD-ROM is a read-only device and data cannot be written to it. However, if you have a special license agreement to copy and repackage the Digital UNIX Version 4.0B operating system, files can be written to the /isl directory of the image, which will be written to the CD-ROM. Refer to Section C.11.4 for more information about burning (writing to) CD-ROMs.

In Digital UNIX Version 4.0, installation cloning could be done only from a network connection to a remote installation services (RIS) server and required user intervention. In Digital UNIX Version 4.0B, however, installation cloning can be done from either a network connection or CD-ROM. In addition, installation cloning can be set up so that it automatically bypasses the following actions that previously required user intervention:


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.2    Overview of Support for User-Supplied Files

The Digital UNIX full installation and installation cloning processes have been enhanced to invoke user-supplied files that contain scripts, programs, or executables to perform user-defined customizations. This ability provides administrators with the opportunity to customize the installation procedure. The files can be provided on diskette, a RIS server, or in the /isl directory of the distribution media (either CD-ROM or an extracted RIS area). Refer to Section C.11.2.1 for things to consider when moving files to an extracted RIS area.

The first invocation of user-supplied files occurs before the actual installation process begins, that is, before any file systems are created and software is installed. At that point, for example, an administrator may want to write a new disk label onto a specific disk to customize disk partitions. This file must be named preinstall.

The second invocation is allowed after software is installed. At that point, for example, an administrator may want to install a customized software application after the installation of the Digital UNIX base software subsets. This file must be named postload.

Refer to Section C.9 and Section C.10 for more information about creating preinstall and postload files for execution during a full installation or installation cloning process.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.3    Relationship Between CDFs and User-Supplied Files

CDFs are used only for an installation cloning process. User-supplied files are invoked and executed during both types of full installations (default and custom) and the installation cloning processes.

CDFs and user-supplied files can be used independently or in any combination. The CDFs and user-supplied files can be located on different sources. For example, the install.cdf file may be on a diskette, the preinstall file might come from the RIS server, and the postload file might come from the /isl directory of the distribution media.

The installation process searches for the install.cdf, preinstall, and postload files in the following order of priority:

  1. The / (root) directory of diskette drive fd0 or fd1. If a diskette is used, it requires a standard UNIX File System (UFS).

  2. The /var/adm/ris/clients/sets/profile_set directory on a RIS server where profile_set is a user-created directory name

  3. The /var/tmp directory on the system to be installed. Keep in mind that CDFs or user-supplied files cannot be delivered in the /var/tmp directory. They can, however, be copied into this directory by executing the preinstall file, which has been previously customized to manipulate a CDF or other user-supplied file.

  4. In the /isl directory of the distribution media (for CD-ROM or RIS installations) or the /isl directory of an extracted RIS area (for RIS installations only).


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.4    Role of the Administrator

To set up a system for installation cloning, an administrator performs the tasks described in Figure C-1. To execute user-supplied files during a full installation, the administrator performs Tasks 3 and 4 only. The numbered list after the task summary describes the tasks in more detail and provides pointers to more information.


Figure C-1: Summary of Administrator Tasks


  1. --> An administrator locates a CDF that is suitable to use for installation cloning. On systems that are installed with Digital UNIX Version 4.0B, the CDF is located in the /var/adm/smlogs directory as the file named install.cdf. There is one CDF generated per system installation. Refer to Section C.6 for a description of the contents of the CDF. Refer to Section C.7 for information about what makes a CDF suitable for installation cloning and for information about acceptable differences between the CDF and the systems to be cloned.

  2. --> The administrator copies and moves the CDF to a working area where it can be optionally modified for installation cloning. The administrator should make a copy of the /var/adm/smlogs/install.cdf file and move and modify the copy. The original CDF should be retained in the /var/adm/smlogs directory because it contains information about the initial system installation that could be valuable for future troubleshooting. The administrator has the option to modify the CDF so that the installation bypasses all user responses usually required during a Digital UNIX installation cloning process. Refer to Section C.8 for information about the attributes in the CDF that can be modified to attain unattended installation cloning.

  3. --> The administrator optionally creates scripts or programs to be executed at two predefined points in the full installation and installation cloning processes. The actions performed by these user-supplied files are determined by the administrator. Refer to Section C.9 and Section C.10 for more information about creating preinstall and postload files for execution during an installation.

  4. --> The administrator moves the modified CDF and any user-supplied files to the / (root) directory on a diskette, to the /var/adm/ris/clients/sets/profile_set directory on a RIS server, or to the /isl directory on a CD-ROM if the Digital UNIX distribution media is being repackaged. The files can also be copied to the /isl directory within an extracted RIS area. Refer to Section C.11 for information about copying the CDF to the appropriate place and the guidelines surrounding each type of distribution media.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.5    Theory of Operation

This section contains a synopsis of how the installation process uses the user-supplied files and CDFs during full and cloned installations. Detailed information is provided in subsequent sections. The work flow shown in Figure C-2 assumes that the administrator has completed the tasks shown in Section C.4.


Figure C-2: Theory of Operation



Figure C-3: Theory of Operation (cont'd)



Figure C-4: Theory of Operation (cont'd)


  1. --> To start an installation process, users boot the system from the Digital UNIX Version 4.0B CD-ROM or over a network connection to a RIS server.

  2. --> The Memory File System (MFS) provides writable space required by the installation process.

  3. --> The installation process searches for a file named preinstall, which is a user-supplied script, program, or executable containing specific actions to be carried out before the installation process begins. If this file is found, it is executed. If execution is successful, the installation process begins. If execution is not successful, the installation process stops. If a preinstall file is not found, the installation process begins the search for a CDF. Refer to Section C.9 for more information about creating a preinstall file.

  4. --> The installation process searches for a CDF that, if found, drives the rest of the installation and begins an installation cloning process. This file, named install.cdf, is searched for in the same order as the preinstall file. If an install.cdf file is not found, a Digital UNIX full installation process continues. Section C.6 provides more information about the CDF.

  5. --> The installation process validates the CDF before beginning an installation cloning process. Validation includes ensuring that the disk name and disk type specified in the CDF exists on the system to be cloned. A CDF validation failure causes the process to stop. For RIS installations, validation includes comparing the versions of the software subsets included in the CDF with the software subset versions that are installed in the RIS environment. Diagnostic messages display the reason for validation failures. Upon successful CDF validation, an installation cloning process continues.

  6. --> The user responses required during a full installation depend upon the type of full installation being performed (default or custom) and the user interface being used (text-based or graphical). Chapter 5 describes the responses required during a full installation process.

  7. --> Upon completion of the software subset load phase, the installation process searches for a file named postload, which is a user-supplied script, program, or executable containing specific actions to be carried out after software subsets are loaded. If this file is found, it is executed. If execution fails, the installation process stops. Refer to Section C.10 for information about creating a postload file.

  8. --> If your system has unattended installation capability, the system automatically reboots after the software subsets are loaded. If your system does not have unattended installation capability, the installation process halts and prompts you to enter commands to reboot the system from the newly installed disks. The screen displays the boot commands that must be entered to reboot the system.

  9. --> The configuration phase begins automatically after the system reboots. Configuration refers to the process of tailoring the software subsets; setting the host name, root password, date, time, geographic location, and time zone; system tuning; and building a kernel. For installation cloning processes, refer to Section C.8.4 about setting these site-specific attributes in the CDF. If values are not defined for these attributes or if the user did not enter a response during the full installation, the installation process becomes interactive to request it.

  10. --> For installation cloning, the type of kernel build is defined in the CDF by the kernel_options= attribute. Refer to Section C.8.3 for the options that are available.

    For full installations, the type of kernel build depends on whether a default or custom installation was performed. Default installations have noninteractive kernel builds that select mandatory kernel options. Custom installations have interactive kernel builds to give users the opportunity to choose the options to build into the kernel.

  11. --> Refer to Table C-5 for information about setting site-specific information if it was not defined in the CDF nor entered during a full installation. If any of these attributes is null, the installation process becomes interactive to request a response from the user.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6    Description of the Configuration Description File

When Digital UNIX Version 4.0B is installed on a system, the installation process creates a configuration description file (CDF). As described previously, the information stored in the CDF can be used to mass-install machines with the same or similar hardware configurations.

The CDF contains the following information about an installation:

The CDF, install.cdf, is located on a newly installed system in the /var/adm/smlogs directory.


NOTE

CDFs used in previous versions of the operating system may not be compatible with Digital UNIX Version 4.0B.


The CDF is in stanza file format, and is logically organized as groupings of attribute-value pairs.
Each attribute-value pair is separated with an equal sign (=). Each logical grouping of attribute-value pairs is defined as an item. Refer to the stanza(4) reference page for more information about stanza file format.

Four items are defined in the installation CDF:


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.1    Sample Configuration Description File

In the sample CDF shown in Example C-1, attributes marked with an asterisk (*) must be manually included into the CDF when it is retrieved from an installed system because the installation interfaces do not currently provide the ability to set these values. Section C.8 defines these attributes and shows you how to include them in the CDF.

Section C.6.2 provides definitions of all attribute-value pairs in the CDF.


Example C-1: Sample Configuration Description File (CDF)
install:
      _item=Inst_islinfo
      prompt=no			*
      media_type=REMOTE
      server_timezone=Eastern
      timeset=1
      server_locality=US
      server=daria
      risdir=/
      _action=create
      srcloc=daria:
      client=kramer

install:
      _item=Inst_filesystem
      maj_min_num=8388608
      disk_number=0
      disk_name=rz0
      controller_type=SCSI
      name=root
      partition=a
      controller_number=0
      disk_type=RZ26L
      _action=create
      file_system_type=UFS

install:
      _item=Inst_filesystem
      maj_min_num=8388608
      disk_number=0
      disk_name=rz0
      controller_type=SCSI
      name=usr
      partition=g
      controller_number=0
      disk_type=RZ26L
      _action=create
      file_system_type=UFS

install:
      _item=Inst_filesystem
      maj_min_num=8388608
      disk_number=0
      disk_name="in /usr"
      controller_type=SCSI
      name=var
      partition=g
      controller_number=0
      disk_type=RZ26L
      _action=create
      file_system_type=UFS

install:
	     _item=Inst_filesystem
	     maj_min_num=8388608
	     disk_number=0
	     disk_name=rz0
	     controller_type=SCSI
	     name=swap1
	     partition=b
	     controller_number=0
	     disk_type=RZ26L
	     _action=create
	     file_system_type=swap

install:
	     _item=Inst_subsets
	     names=OSFBASE410,OSFBIN410,OSFBINCOM410,OSFCDEDT410,OSFCDEMAIL410,
OSF CDEMIN410,OSFCLINET410,OSFCMPLRS410,OSFDPSFONT410,OSFFONT15410,OSFHWB
ASE410,OSFHWBIN410,OSFHWBINCOM410,OSFKBDLK401410,OSFMITFONT410,OSFNETCON
F410,OSFNETSCAPE410,OSFNFS410,OSFNFSCONF410,OSFOLDX11410,OSFPRINT410,OSF
SER410,OSFSERTC410,OSFSYSMAN410,OSFTCLBASE410,OSFTKBASE410,OSFX11410,OSF
XADMIN410,OSFXPRINT410,OSFXSYSMAN410
	     _action=create
	     advflag=1

install:
	     _item=Inst_cinstall
	     kernel_option=all			*
	     password=C36V.nMSW0j/o
	     timeset=yes
	     timezone=Eastern
	     locality=US
	     _action=create
	     hostname=kramer


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.2    Attribute-Value Pair Definitions

This section provides definitions for all attribute-value pairs in the CDF.

The attribute-value pairs within individual items differ as a result of the distribution method (CD-ROM or RIS) that was used to perform the initial installation of the model system.


Caution

Digital recommends that only experienced system administrators modify the attributes-value pairs in the CDF. If you are an experienced administrator, Digital does not recommend editing the CDF other than for those attribute-value pairs in the Inst_cinstall item and those marked with an asterisk in the sample CDF shown in
Example C-1. Typographical errors and inserting attribute-value pairs into the incorrect item may result in serious corruption on the cloned systems and may render the systems unusable.

In addition, attribute-value pairs cannot contain blank spaces. Blank spaces cause data validation errors. Be very careful to remove all blank spaces especially at the end of a line. When you want to give an attribute a null value, make sure there is nothing (null) after the equal sign (=).

Do not modify or remove attributes that are prefixed with an underscore (_). These attributes, for example _action=create, are internal variables required by the full installation and installation cloning processes.



[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.2.1    Attributes in the Inst_islinfo Item

Table C-2 defines the attributes in the Inst_islinfo item in the CDF. The Inst_islinfo item is used to convey the system state before the start of the installation process.
Table C-2: Attribute Definitions in the Initial Subset Load (Inst_islinfo) Item
AttributeDefinition
client=  This attribute is valid only for RIS full installations (not installation cloning) and specifies the client name of the system that was cloned. The client name is automatically determined as a result of the bootp request to the server. Do not modify this attribute for installation cloning because the value in this attribute does not have to match the client systems to be cloned. 
clone=  This attribute is automatically inserted into the CDF as a result of an installation cloning process and is only valid during the installation cloning process. This attribute-value pair should not be manually set. 
media_type=  This attribute is used by the full installation and installation cloning processes to indicate the type of distribution media for the current installation. This is the only required entry in the Inst_islinfo item. Valid values are REMOTE and CDROM. Edit this attribute when the type of distribution media used for the initial installation is different than the installation cloning that is to take place. 
prompt=  This attribute is used by the installation cloning process to indicate whether the start of an installation cloning process requires a confirmation response from the user. This attribute must be manually entered into the CDF for an installation cloning process because the installation interfaces do not provide the ability to insert this attribute into the CDF.A value of yes indicates that the process should prompt for confirmation to use the CDF. A value of no indicates that the installation cloning process should use this CDF and bypass the confirmation question. If this attribute is not included in the CDF, the default is prompt=yes. Setting the attribute to no should be used with caution because the installation cloning begins as soon as the installation process detects a CDF. If you wanted to boot the system from the distribution media and perform system management or disk maintenance tasks, for example, you would not want the installation cloning to begin immediately. 
risdir=  This attribute is specific to RIS full installations and is automatically set to the base RIS directory of the product environment to which the client system is registered. Do not modify this attribute for installation cloning. 
server=  This attribute is specific to RIS full and cloning installations and identifies the RIS server to which the client system is currently registered. Do not modify this attribute for installation cloning. 
server_locality=  This attribute is specific to RIS full installations and specifies to the installation interfaces the current geographic location. Do not modify this attribute for installation cloning. 
server_timezone=  This attribute is specific to RIS full installations and specifies to the installation interfaces the current geographic timezone. This value is automatically set during a RIS full installation. Do not modify this attribute for installation cloning. 
srcloc=  This attribute is not used by either the full installation or installation cloning processes; it is used by the operating system for internal purposes. This attribute identifies the location of the software to load. For RIS installations, this value specifies the server name (appended with a colon). For CD-ROM installations, this value is the directory path /ALPHA/BASE. Do not modify this attribute unless the media_type attribute is changed because this value must be consistent with the value of media_type
timeset=  This attribute applies to full installations and indicates to the installation interfaces whether the date and time on the client system have been successfully set and whether the date and time can be displayed during the installation. Valid values are: 0- Date and time have not been set and will not be displayed during the installation process 1- Date and time have been successfully set and will be displayed where appropriate during the installation process Do not modify this attribute for installation cloning. 


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.2.2    Attributes in the Inst_filesystem Item

Table C-3 defines the attributes in the Inst_filesystem item in the CDF. The Inst_filesystem item is used to convey information about the number and type of file systems that are to be created on the cloned system. At a minimum, there must be at least four file system items to describe the / (root), /usr, and /var file systems and one swap area. Except where noted, you optionally can modify all attribute-value pairs in this item, although Digital does not recommend editing the CDF.
Table C-3: Attribute Definitions in the File System (Inst_filesystem) Item
AttributeDefinition
name=  This attribute is a required attribute that specifies the name of the file system to be made. Valid values are: root, usr, var, swap1, and swap2. There can only be one item each for root, usr, var, swap1, and swap2
file_system_type=  This attribute is a required attribute that specifies the file system type to be created for the named file system. Valid values are: ufs, advfs, and swap. If the value of the name= attribute is swap1 or swap2, the value of this attribute must be swap.


Caution

Be aware that changing this value from ufs to advfs may cause errors on the cloned system because the software subsets necessary to support an Advanced File System (AdvFS) may not be defined in the CDF and will not be installed on the cloned system. Therefore, the file system will be unreadable. Do not change this value to advfs unless other file systems have been set by the installation process to advfs or the required AdvFS software subsets are present in the names= attribute in the Inst_subsets item.


 

disk_name=  This attribute is a required attribute that specifies the disk name for the named file system as it is known to the operating system (for example, rz0). The value in this attribute must be consistent with (or match) the value in the disk_type= attribute. If you change this attribute, you must validate the change with respect to the disk_type= attribute. For example, if you change this value to disk_name=rz1, you must determine the type of disk at rz1. If it is an RZ58 type of disk, make sure the value of the disk_type= attribute is RZ58
disk_type=  This attribute is a required attribute that indicates the type of disk for the specified disk_name (for example RZ26). The value in this attribute must be consistent with the disk_name= attribute. Refer to the disk_name= attribute for more information. 
partition=  This attribute is a required attribute that specifies the disk partition on which the named file system will be created. Valid values are the letters a through h inclusive. The root file system must always be located on partition a. If you change the value in this attribute for any file system other than root, make sure the partition you choose does not overlap another partition. 
controller_type=  This attribute identifies the controller type to which the specified disk for the named file system is connected. During a full installation, this value is automatically provided for informational purposes. During an installation cloning process this attribute is not used, and can be omitted from the CDF. 
controller_number=  This attribute identifies the controller number to which the specified disk for the named file system is connected. During a full installation, this value is automatically provided for informational purposes. During an installation cloning process this attribute is not used, and can be omitted from the CDF. 
maj_min_num=  This value is automatically calculated for full and cloned installations so there is no need to modify it. This attribute is required for the root file system item and specifies the major and minor number of the specified disk for the named file system. The major and minor number is used to map the software device name (as known to the operating system) to the firmware device name (as known to the SRM console) so that the proper boot commands are displayed on the screen during the manual boot phase of the installation. 


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.2.3    Attributes in the Inst_subsets Item

Table C-4 defines the attributes in the Inst_subsets item in the CDF. The Inst_subsets item is used to convey information to the installation cloning process about the Digital UNIX base software subsets that are to be installed on the system to be cloned.
Table C-4: Attribute Definitions in the Software Subsets Load (Inst_subsets) Item
AttributeDefinition
advflag=  Digital recommends that you do not modify this attribute. This attribute is a required attribute that specifies the type of installation (custom or default) that is to occur. Valid values are: 
  0- Default installation 1- Custom installation 
 

Caution

Be aware that changing the value of this attribute may cause the setld command to fail during software subset loading because the software subsets defined in the CDF may not be compatible with the type of installation defined by this attribute.


 
  Setting this attribute to 0 nullifies the kernel_option= attribute in the Inst_cinstall item because default installations provide noninteractive kernel builds with mandatory kernel options. 
names=  This attribute is a required attribute that specifies the list of Digital UNIX base software subsets to be installed. Each software subset name is separated by a comma (,) and must be on one continuous line (let the line wrap). If you add software subset names to this attribute, you must consider available disk space and dependencies upon other software subsets. Refer to Appendix D for software subset dependency information and disk space requirements. 


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.6.2.4    Attributes in the Inst_cinstall Item

Table C-5 defines the attributes in the Inst_cinstall item in the CDF. The Inst_cinstall item is used to convey client system configuration information to the installation cloning process. All of the attributes specified in the installation configuration item are optional. If values are not provided for these attributes, the installation process becomes interactive to request this information during the installation configuration phase.

To use a single CDF to clone many systems, consider leaving the system-specific attributes such as host name and password null, but provide attributes for site-specific attributes such as kernel option, time zone, geographic location, and date and time.


Table C-5: Attribute Definitions in the Installation Configuration (Inst_cinstall) Item
AttributeDefinition
hostname=  This attribute specifies the client system's host name to the installation process. Host names for client systems that exist on the same network must be unique. Refer to the Installation Guide for guidelines on choosing a proper host name. During a RIS installation cloning process, this value is automatically set to the host name of the client system. For CD-ROM installations, make sure this value is set correctly or is null. A null value means that the installation process becomes interactive during the installation configuration phase to request a host name. 
kernel_option=  This attribute specifies to the installation process whether the tailored kernel build should be interactive or noninteractive. 
  This attribute must be manually entered in the CDF for an installation cloning process because the installation interfaces do not provide the ability to insert this attribute in the CDF. 
  In an interactive kernel build session, a kernel options menu is presented allowing selection of any or all optional kernel options. To specify an interactive tailored kernel build, use the following value:  
  kernel_option=interactive 
  For noninteractive kernel builds, two options are provided:  
  kernel_option=mandatory  
  kernel_option=all 
  The mandatory value builds a tailored kernel with only mandatory kernel options. The all value builds a tailored kernel with all mandatory and optional kernel options. 
  The default behavior of a full, custom installation is the interactive type of kernel build. Full, default installations have mandatory type kernel builds. 
  If the value of the advflag attribute in the Inst_subsets item is zero (0), the value given to the kernel_option attribute value is ignored. 
locality=  This attribute specifies the geographic location of the client system. Valid values for this attribute are located on an installed system in the /etc/zoneinfo directory, which contains an entry (a file or a directory) for each geographic location. During a RIS installation cloning process, this value is automatically set to the geographic location of the RIS server. A null value means that the installation process becomes interactive during the installation configuration phase to request a geographic location. 
password=  This attribute specifies to the installation process the encrypted root password for the client system. The presence of a value here means that all cloned systems share the same root password. A null value means that the installation process becomes interactive during the installation configuration phase to request a password. 
  Because the value of password= must be encrypted, you cannot manually enter a new value for this attribute. 
timeset=  This attribute specifies to the installation process that the system date and time have already been set on the client system. In the case of a RIS full installation or RIS installation cloning, this value is always set to yes. Valid values are:  
  no- System date and time have not been set. The installation process becomes interactive to request the date and time.  
  yes- System date and time have been set. For CD-ROM installations, users should verify the accuracy of the date and time after logging in for the first time because the installation process may not have set it correctly. 
timezone=  This attribute specifies the time zone within a specific geographic location (if applicable). Valid values for this attribute are located in the subdirectories of the /etc/zoneinfo directory. During a RIS installation cloning process, this value is automatically set to the time zone of the RIS server. The value of timezone must be a valid time zone for the geographic location defined in the locality= attribute. For example, if locality=US, only time zones in the United States are valid. If the geographic location does not have a time zone, leave this value null. The installation process recognizes geographic locations that do not have time zones, and will not request a time zone during the configuration phase. 
  If the geographic location has valid time zones, a null value means that the installation process becomes interactive during the installation configuration phase to request a time zone. 


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.7    Generating or Selecting an Appropriate CDF

When generating a CDF through the installation of a system or selecting which CDF to use to clone other similar systems, you must consider the disk configuration, graphics adapter, font sizes and keyboard types of the systems to be cloned. Ideally, however, you should clone systems with identical hardware configurations.

To reduce the disk space required when Digital UNIX is installed, the software required to support the different graphics adapters, font sizes, and keyboard types has been packaged so that only the software subsets required to support options present on the system are mandatory and installed automatically. All other software subsets are considered optional and are not installed unless you specifically select them. Determining the mandatory software subsets for a system is done automatically by the installation process and guarantees that only appropriate software subsets are installed.

However, when a system is installed using installation cloning, the software subsets installed on to the system are defined in the CDF. Therefore, if the system to be cloned has a different graphics adapter, font size, or keyboard type than the system on which the CDF was created, the appropriate software subsets will not be installed and the cloned system may not be usable.

To generate a CDF that is versatile enough for use across differing systems, you may want to consider installing a system to use as a model. That is, perform a custom installation on a model system so that the CDF generated from that installation is usable by systems with different graphics adapters, font sizes, and keyboards. You do this by installing the software subsets to support all graphics adapters, font sizes, and keyboard types required by the systems to be cloned even though they are not required by the model system.

Acceptable differences in disk configuration, graphics adapter, font sizes, and keyboard type are explained in the following sections.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.7.1    Acceptable Differences in Disk Configurations

The system to be installed by the installation cloning process should have the same hardware configuration as the system where the CDF was generated. However, it is possible to support slight differences in configuration.

The system to be cloned must have the same disk configuration for the disks on which root, usr, swap1, var (if it is not a directory under /usr) and swap2 (if allocated) are to be installed as the system on which the CDF was generated. The same disk configuration means that the disk type (for example RZ26) and the device name (for example rz0) must match. If the partition tables for these disks are not identical on both systems, the software defined in the CDF may not fit on to the system to be cloned or would overlap the disk partitions.


Note

You may want to consider writing a preinstall script to install a common disk label on all systems to be cloned.
Example C-2 contains a sample script that installs a common disk label.


It does not matter if disks other than those used for the file systems and swap areas created during an installation are different on the system to be cloned.

Table C-6 illustrates acceptable differences in disk configuration between a CDF generated from a model system and a system to be cloned.


Table C-6: Acceptable Differences in Disk Configuration Between a Model System and a System to be Cloned
SystemDisk TypeDevice Name
model system  RZ26RZ25  rz0 °rz1 
system to be cloned  RZ26RZ26  rz0rz1 

Assuming there are no other differences in disk configuration, the system to be cloned can use the CDF generated from the model system. The difference in disk type at device name rz1 is acceptable because the file systems and swap space were not placed on it. If the disk device at rz0 were different, however, an installation cloning could not be performed.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.7.2    Considering Differences in Graphics Adapters

When you install a model system from which you will use the CDF to clone other systems, you must consider the graphics options of the systems that will be cloned. If any of the systems to be cloned have different graphics options, the software subsets required to support the graphics options needed by those systems must be installed on the model system.

When selecting software subsets, look in the Windowing Environment category for software subsets starting with the words X Servers for <name>. Replace <name> with the name that describes the graphics options supported by the software subset. In Digital UNIX Version 4.0B, the following graphics software subsets are available:


Note

X Servers for PCbus adapters supported by Digital UNIX are specified in the Software Product Description (SPD).


Table C-7 displays the graphics adapters on a model system and a system to be cloned. The hardware configuration of the model system and the system to be cloned are determined to be similar enough to allow the CDF from the model system to be used for the installation cloning.


Table C-7: Acceptable Differences in Graphics Adapters Between a Model System and a System to be Cloned
SystemGraphics Adapter
model system  Open3D 
system to be cloned  QVision (PCbus) 

During the installation of the model system, the X Servers for Open3D software subset is considered mandatory for the model system and is automatically installed. The X Servers for PCbus software subset is considered optional for the model system. Installing this optional software subset on the model system ensures that the appropriate software is available for the system to be cloned. If you do not install the X Servers for PCbus onto the model system, the graphics capabilities of the system to be cloned are likely to be disabled.

Caution

Do not use the CDF from a system that does not have graphics capabilities to clone systems that have the hardware to support graphics. There are several software subsets, most notably those associated with the common desktop environment (CDE), that will not be loaded on systems without graphics capabilities that are mandatory for systems with graphics capabilities. If you use a CDF from a system without graphics capabilities to clone a system with graphics capabilities, the desktop environment on the cloned system will be corrupted.


If you are unsure of which graphics options are available on the systems you want to clone, install all of the graphics software subsets that are available. However, installing all of the software subsets requires more disk space than loading only selected graphics software subsets.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.7.3    Considering Differences in Font Size

To reduce the disk space required when Digital UNIX is installed, the software required to support the 75dpi (dots per inch) and 100dpi font sizes are contained in separate software subsets.

During an installation cloning, the font software subsets to be installed are defined in the CDF. If the system to be cloned requires a different size font than those defined by the software subsets in the CDF, the system to be cloned will not have the appropriate fonts loaded.

When generating the CDF through the full installation of a model system, you must consider the font sizes required by the systems to be cloned from the CDF. If the systems to be cloned require different size fonts, load the appropriate font software subset when installing the model system.

The need for DECwindows 75dpi Fonts or DECwindows 100dpi Fonts depends on the resolution of the graphics adapter being used. On a system already installed with Digital UNIX, this value can be determined by entering the following command:

# sizer -gr
When the resolution is 1024x768 or less, the DECwindows 75dpi Fonts are required. When the resolution is greater, the DECwindows 100dpi Fonts are required. If you are unsure of the resolution available on the systems to be cloned, select both font software subsets to ensure that the correct font is available.

Systems with multiple graphics adapters may require both the DECwindows 75dpi Fonts and DECwindows 100dpi Fonts if the adapters include those with 1024x768 or less resolution and those with greater resolution.

While there are other software subsets that contain fonts, only the DECwindows fonts are packaged separately by size.

Table C-8 displays the different font sizes required on a model system and a system to be cloned. The hardware configuration of the model system and the system to be cloned are determined to be similar enough to allow the CDF from the model system to be used for the installation cloning.


Table C-8: Acceptable Differences in Font Sizes Between a Model System and a System to be Cloned
SystemGraphics ResolutionRequired Font Size
model system  1024x680  DECwindows 75dpi Fonts 
system to be cloned  1280x1024  DECwindows 100dpi Fonts 

During the installation of the model system, the DECwindows 75dpi Fonts software subset is mandatory and is installed automatically; the DECwindows 100dpi Fonts software subset is optional. You should install the optional software subset to provide the necessary fonts for the installation cloning of the client system.

If you are unsure of the fonts available on the systems you want to clone, you can ensure that you provide the appropriate fonts by installing all of the font software subsets on to the model system. Installing all of the font software subsets will require more space than loading selected fonts.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.7.4    Considering Differences in Keyboard Type

To reduce the disk space required when Digital UNIX is installed, the software subsets required to support the different Digital keyboard types is contained in separate software subsets.

During an installation cloning, the keyboard support software subset to be installed is defined in the CDF. If the system to be cloned has a different keyboard type than the model system, the cloned system will not have the appropriate keyboard software installed.

When generating the CDF through the installation of a model system, you must consider the keyboard type of the systems that will be cloned using the CDF. If the systems that will be cloned have different keyboard types, load the appropriate keyboard support software subset when installing the model system. The keyboard type can be determined from information available when the system is in console mode or by looking at the model number on the underside of the keyboard.

Table C-9 displays the keyboard types on a model system and a system to be cloned. The hardware configuration of the model system and the system to be cloned are determined to be similar enough to allow the CDF from the model system to be used for the installation cloning.


Table C-9: Acceptable Differences in Keyboard Types Between a Model System and a System to be Cloned
SystemKeyboard Type
model system  PXCAL 
system to be cloned  LK444 

During the installation of the model system, the software subset PCXAL Keyboard Support is mandatory and is installed automatically. The software subset for LK444 Keyboard Support is optional. Selecting this optional software subset results in some unnecessary software being loaded on the model system but allows the CDF to be appropriate to clone the client system.

If you are unsure of the keyboard types available on the systems you want to clone, you can ensure that you provide the appropriate keyboard type by installing all of the keyboard software subsets. However, loading all keyboard software subsets will require more disk space than loading selected keyboard software subsets.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.8    Modifying Attributes in the CDF to Achieve Unattended Installations

Digital recommends that only experienced system administrators modify the attributes-value pairs in the CDF. Before modifying the CDF, make sure you read the information in the Caution in Section C.6.2.

Do not modify the original CDF located in the /var/adm/smlogs directory of an installed system. Instead, make a copy of install.cdf and modify the copy. The original install.cdf file contains information related to the system installation that could be valuable for future use. You should retain the install.cdf file in the /var/adm/smlogs directory.

Some attribute-value pairs must be manually added to the CDF for an installation cloning process because the installation interfaces do not currently provide the ability to set these values. The following sections describe the attribute-values pairs that can be manually added to the CDF to attain unattended installations.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.8.1    Errors in the CDF

While modifying a CDF, a common error is to include a trailing blank space after an attribute-value pair. If the validation process detects a trailing blank space in the CDF, a message similar to the following will be displayed:
- ----------------------------------------
Some errors occurred:
SetItmAttr: invalid attribute value kernel_option=all
- ----------------------------------------
This error causes the installation process to stop. In the previous example, the validation process found a trailing blank space after the word all in the kernel_option=all attribute-value pair. The corrective action is to edit the CDF and remove the blank space. Then, restart the installation process at the client system.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.8.2    Modifying the CDF Confirmation Attribute

Previous versions of the installation cloning process required the user to confirm that the CDF was to be used to start an installation cloning rather than a full installation. The purpose of this confirmation was to protect a system from an inadvertent installation cloning if the system was mistakenly still registered to a RIS environment and CDF.

The CDF confirmation question is now configurable through the prompt= attribute-value pair in the Inst_islinfo item in the CDF. The value of the prompt= attribute determines whether confirmation is required before the CDF is used to start an installation cloning process. Valid values are:

If this attribute-value pair is not defined or is null, the installation cloning process defaults to prompt=yes.

A portion of a CDF in the following example shows you where to include the prompt= attribute-value pair in the Inst_islinfo item:

install:

_item=Inst_islinfo
prompt=no
media_type=CDROM
server=cosmo
_action=create
srcloc=/ALPHA/BASE


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.8.3    Modifying the Tailored Kernel Build Attribute

A Digital UNIX default installation provides a noninteractive kernel build with mandatory kernel options enabled. A custom installation provides an interactive kernel build and allows you to tailor the kernel by allowing you to select mandatory and optional kernel options.

The kernel_option attribute in the Inst_cinstall item allows a noninteractive tailored kernel build with all kernel options (mandatory and optional) or mandatory kernel options only. In addition, the interactive value can be specified to allow you to tailor the kernel. The values for the kernel_option attribute are defined as follows:

A portion of a CDF in the following example shows you where to include the attribute-value pair into the Inst_cinstall item:

install:

_item=Inst_cinstall
kernel_option=all
password=SdDt78fuPrMkE
timeset=yes
timezone=Eastern
locality=US
_action=create
hostname=kramer

Kernel build failures that occur during a noninteractive kernel build cause the kernel build process to become interactive and provides the user with options for proceeding.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.8.4    Modifying Site- and System-Specific Attributes

You must read this section if you plan to perform installation cloning from CD-ROM.

Setting site- and system-specific information such as host name, geographic location, time zone, date, and time are trivial in the case of a RIS installation because these values are obtained from the RIS server automatically during the installation. This statement is true for full installations from RIS or from a RIS installation cloning process.

In the case of a standalone system installed by a CD-ROM installation cloning process, however, setting these values must be determined from the CDF that drives the installation cloning. If the CDF does not define these attributes, the values must be entered interactively during the configuration phase of the installation cloning process that occurs after software has been loaded.

The system-specific attributes to be considered are:

The site-specific attributes to be considered are:


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.9    Creating preinstall Files

The installation process tests for the existence of customer supplied files at predefined invocation points. The first invocation point is between the creation of the memory file systems (MFS) and the search for a CDF. At this point, the installation process searches for a file named preinstall, which is a user-supplied script, program, or executable containing specific actions to be carried out before the file system creation and software subset load phases of the installation process.

Actions to be carried out before file systems are created and software subsets are loaded might include writing a customized disk label to one or more disks.

You would not want the preinstall file to execute any function that requires the installed file systems and software to be available because these phases of the installation have not yet been completed.

The user-supplied file must be named preinstall, and the preinstall file and any files that it calls require execute permission.

It is not necessary that this file be contained in the same location in which the CDF and postload files are found.

If execution of the preinstall file fails, the preinstall file is responsible for supplying its own status or error messages. Digital does not guarantee the results of executing the script or program but does guarantee that upon successful completion, the installation process proceeds.

The installation process queries the return status from the execution of the preinstall file and terminates the installation process if a non-zero return status is received.

The installation process searches for the preinstall file in the following order of priority:

  1. The / (root) directory of diskette drive fd0 or fd1. If a diskette is used, it requires a standard UNIX File System (UFS).

  2. The /var/adm/ris/clients/sets/profile_set directory on the RIS server. Profile set directories are created by the RIS or system administrator. Refer to Section C.11.2 for more information about profile set directories on RIS servers.

  3. In the /isl directory of the distribution media or to the /isl directory of an extracted RIS area.

The sample preinstall script shown in the following example applies a customized disk label to an RZ26 disk.


Example C-2: Sample preinstall Script
#!/sbin/sh

#
# Write a custom disk label to the
# system disk before starting the installation.
#

# NOTE:  THIS FILE ASSUMES A DISK NAME OF rz0 AND DISK TYPE OF RZ26

#
# Make the device special file for rz0
#
(cd /dev; ./MAKEDEV rz0)

#
# First, zero the label
#
2>/dev/null disklabel -z rz0

#
# Next, restore the label
#
disklabel -Rr rz0 ./DLSAVE RZ26 ||                      [1]
{
	echo "\nError restoring disklabel on rz0\n"
	exit 1
}

echo "\nThe disklabel that has been applied is:\n"
disklabel -r rz0 | tail -10
exit 0

  1. --> The DLSAVE file called by the preinstall script must reside in the same directory as the preinstall script.

The sample DLSAVE file required by the preinstall script is shown in Example C-3. The DLSAVE file contains a disk label that was created by reading the disk label of the disk at rz0 and redirecting the output into a file. To create this file, you would enter commands similar to the following:

# disklabel -r rz0 > DLSAVE


Example C-3: DLSAVE File Required By the Sample preinstall Script
# /dev/rrz8a:
type: SCSI
disk: rz26
label:
flags:
bytes/sector: 512
sectors/track: 57
tracks/cylinder: 14
sectors/cylinder: 798
cylinders: 2570
sectors/unit: 2050860
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0		# milliseconds
track-to-track seek: 0	# milliseconds
drivedata: 0

8 partitions:
#      size  offset  fstype [fsize bsize  cpg]
 a:  131072       0  4.2BSD   1024  8192   16  # (Cyl.    0 - 164*)
 b:  262144  131072  unused   1024  8192       # (Cyl.  164*- 492*)
 c: 2050860       0  unused   1024  8192       # (Cyl.    0 - 2569)
 d:  552548  393216  unused   1024  8192       # (Cyl.  492*- 1185*)
 e:  552548  945764  unused   1024  8192       # (Cyl. 1185*- 1877*)
 f:  552548 1498312  unused   1024  8192       # (Cyl. 1877*- 2569*)
 g: 1210000  393216  4.2BSD   1024  8192   16  # (Cyl.  492*- 2009*)
 h:  447644 1603216  4.2BSD   1024  8192   16  # (Cyl. 2009*- 2569*)


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.10    Creating postload Files

Upon completion of the file system creation and software subset load phases and the preparation of the configuration environment for the pending configuration phase, the installation process searches for a file named postload, which contains specific actions to be carried out.

Actions to be carried out after software subsets are loaded might include creating additional file systems or installing additional software that was not installed as part of the Digital UNIX base operating system.

The postload file and any files that postload calls require execute permission. The installation process searches for the postload file in the following order of priority:

  1. The / (root) directory of diskette drive fd0 or fd1. If a diskette is used, it requires a standard UNIX File System (UFS).

  2. The /var/adm/ris/clients/sets/profile_set directory on the RIS server. Profile set directories are created by the RIS or system administrator. Refer to Section C.11.2 for more information about profile set directories on RIS servers.

  3. In the /var/tmp directory on the system to be installed.

  4. In the /isl directory of the distribution media or to the /isl directory of an extracted RIS area.

It is not necessary that the postload file be contained on the same media on which the CDF and preinstall file are found.

The installation process queries the results of the execution of the postload file and terminates the installation process upon a non-zero return status.

It is important to know that at this invocation point, the newly created root, /usr, and /var file systems on the magnetic media are mount-relative with respect to the directory /mnt until the system is rebooted from the default boot device. That is, the root file system is /mnt, the usr file system is /mnt/usr, and so on.

The sample postload script shown in Example C-4 is creating a new file system called users and is then adding the entry into the /etc/fstab file to mount the new file system upon every reboot.


Example C-4: Sample postload Script
#!/sbin/sh
#
# postload - script which is invoked after the subset load of a full
#	installation.  The script creates a new file system and
#	adds an entry in the fstab file.  Doing this will make the
#	file system available as soon as the installation completes.

#
# Create a new file system on rz2c which is to be mounted at /usr/users
#

echo "postload:  creating new file system on rz2c\n"

# First, make sure that all device special files exist

(cd /dev; ./MAKEDEV rz2)

# Next, create the UFS file system on rz2c, an RZ26L disk.

/usr/sbin/newfs -F /dev/rrz2c RZ26L ||
{
	   echo "postload:  failed to create a new file system on rz2c\n"

	   # We consider this a nonfatal error and allow the install to
	   # continue.  This is done by returning 0.  Otherwise, exit with a
	   # non-zero value.

	   exit 0
}

# Next, add an entry to fstab so that this new file system is
# automatically mounted when the system boots.

# NOTE:  the actual installed file systems are mounted at /mnt.
# Therefore, we want to add the entry to /mnt/etc/fstab and
# not /etc/fstab.

echo "/dev/rz2c	/usr/users	ufs rw 1 2" >> /mnt/etc/fstab

# Finally, make sure the mount point is created.  Once again, create it
# relative to /mnt.

/bin/mkdir /mnt/usr/users

# Process complete!

exit 0


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11    Moving the CDF and Files to the Appropriate Destination

It is the administrator's responsibility to place the install.cdf file, the preinstall and postload files and all files required by preinstall and postload into the appropriate directories so the installation process can find them. Depending upon how you want to deliver the CDF and all related files, you can copy them to the following destinations:

During an installation cloning, the cloning process searches for the CDF and user-supplied files in the following order of priority:

  1. Diskette drive fd0 or fd1

  2. The /var/adm/ris/clients/sets/profile_set subdirectory on the RIS server

  3. The /var/tmp directory on the system to be installed.

  4. The /isl directory on the distribution media (local CD-ROM or extracted RIS area). Refer to Section C.11.2.1 for things to consider when moving files to an extracted RIS area.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.1    Moving the CDF and Files to a Diskette

Before you can copy the CDF and user-supplied files to the diskette, you must first format the diskette, write a new disk label, and then create a new file system using the following command syntax:

fddisk-fmt raw_diskette_device

disklabel-wr diskette_drive disk_type

newfsraw_diskette_device_partition

Use commands similar to the following to format the diskette in diskette drive fd0, write a new disk label specifying the rx23 type of diskette, and creating a new file system on the entire diskette (partition c):

  1. Enter commands similar to the following to format a diskette drive fd0:
    # fddisk -fmt /dev/rfd0

  2. Enter commands similar to the following to write a new disk label to an rx23 type of diskette. The diskette type is printed on the diskette.
    # disklabel -wr fd0 rx23

  3. Use commands similar to the following to create a new file system on the entire diskette, the c partition:
    # newfs /dev/rfd0c

If either the preinstall or postload files are located on the diskette, all files called by the preinstall or postload files must be located on the diskette.

Use commands similar to the following to mount the diskette drive and copy the CDF and all related files to the diskette:

  1. Mount the diskette drive on the /mnt mount point:
    # mount /dev/fd0c /mnt

  2. Enter the chmod command to ensure all files have execute permissions:
    # chmod 777 *
    The asterisk (*) is a wildcard character that represents all files in the directory.

  3. Assuming that you are in the directory in which the files are located, enter the following copy commands to copy the files to the diskette:
    # cp ./install.cdf /mnt/install.cdf 
    
    # cp ./preinstall /mnt/preinstall 
    
    # cp ./postload /mnt/postload 
    
    # cp ./file_name /mnt/file_name

  4. Unmount the diskette drive:
    # umount /mnt 


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2    Moving the CDF and Files to a RIS Server

The information contained in this section applies to RIS servers running Digital UNIX Version 4.0A and later. This functionality was different on RIS servers running Digital UNIX Version 4.0. For information about moving the CDF and user-supplied files to a RIS server running Version 4.0, see the appropriate Digital UNIX Version 4.0 documentation.

The Remote Installation Services utility (RIS) has been modified to support client registration to both RIS environments and profile set directories. RIS maintains the CDFs and user-supplied files in logically organized subdirectories that are created by the RIS administrator. These subdirectories, known as profile sets must be located within the /var/adm/ris/clients/sets directory. The administrator uses the mkdir command to make profile set directories.

A profile set is a directory that contains the files used during an installation process. The sets directory can contain many profile sets. Each of the profile set directories may contain a CDF (install.cdf), a preinstallation file (preinstall), a postinstallation file (postload), and all files called by the preinstall and postload files. All files are optional; they can be used independently or in any combination. It is the RIS administrator's responsibility to place the appropriate files into the correct profile set directory.

The profile_set directories you create depend upon your working environment and how you want to logically organize the functions of the CDFs and files. If, for example, your site or facility requires engineering workstations to be installed and configured different than the workstations in the accounting department, you might want to create two profile set directories; one named engineering and one named accounting. Those profile sets would contain the CDFs and files that were created to suit the configuration needs of both departments.

Another hypothetical situation for defining profile sets is one in which separate CDFs and files are maintained for server type systems and workstation type systems. Profile set directories named server and workstation might be set up under that scenario.

Use procedures similar to the following to copy the CDF, preinstall and postload files, and related files to a profile set directory:

  1. Change to the /var/adm/ris/clients/sets directory, and using the naming scheme of your choice, create a profile set directory with an appropriate name:
    # cd /var/adm/ris/clients/sets
    
    # mkdir engineering

  2. To ensure files are copied to the correct directory, change to the new profile set directory:
    # cd engineering

  3. Using the copy tool you usually use (for example, ftp, dcp, or ,rcp) copy the modified CDF and optionally the preinstall, postload, and all other related files from your working area to the new engineering profile set directory.

  4. Enter the chmod command to ensure all files have execute permissions:
    # chmod 755 *
    The asterisk (*) is a wildcard character that represents all files in the directory.

After you copy the appropriate CDF and other files to the profile sets directory, you can register RIS clients for installation cloning or for user-defined file invocation during a full RIS installation. You do this by registering new clients to a RIS environment as well as to a profile set. If a RIS client is registered to a profile set and boots across the network to start an installation, the order of priority in which a search for a CDF and other optional files is done is shown in Section C.11. If a CDF is found, it is retrieved and used by the installation process to provide the answers to all installation configuration questions.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.1    Moving Files to an Extracted RIS Area

If an install.cdf, preinstall, or postload file is moved to the /isl area of an extracted RIS area, the files will be used by all client systems installing from that RIS area.

If this action is not appropriate, the administrator should create profile set directories to supply these files on a client-by-client basis.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.2    Changes to the RIS Interface

The following changes have been made to the Digital UNIX Version 4.0B RIS interface to accommodate the addition of profile set directories:


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.3    Registering a RIS Client to a Profile Set

Follow the general procedures in Sharing Software on a Local Area Network to register a client system to a RIS environment and a profile set. The Sharing Software on a Local Area Network guide was not updated to reflect the new prompts shown in Section C.11.2.2.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.4    Determining Registration for RIS Clients

To determine if a RIS client is registered to a profile set, examine the RIS database file, /var/adm/ris/clients/risdb, on the RIS server. The name of the profile set is specified in the fourth field; fields are separated by a colon. In the following sample entry in the risdb file, the client system kramer is registered to the engineering profile set:
kramer:08-00-2b-58-89-1c:ris2.alpha,product_1:engineering


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.5    Removing a RIS Client from Profile Set Registration

You can remove a client from profile set registration by using the Modify option from the RIS Utility Main Menu. When you are prompted to specify a profile set for the client, enter n or press Return to register the client without specifying a profile set.


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.2.6    Deleting Profile Sets from the RIS Server

If a profile set is no longer needed, you can delete it by removing the appropriate profile_set directory from the directory /var/adm/ris/clients/sets.

Examine the RIS database file on the RIS server, /var/adm/ris/clients/risdb, before deleting a profile set to ensure that no clients are registered to it. The name of the profile set is specified in the fourth field; fields are separated by a colon (:). In the following sample entry in the risdb file, the client newman is registered to the accounting profile set:

newman:08-00-2b-58-89-1c:ris2.alpha,product_1:accounting


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.3    Moving the CDF and Files to the /var/tmp Directory

The /var/tmp directory is a writable directory created during the installation process and, therefore, cannot be used to ship the CDF and user-supplied files. However, if a preinstall script is used, it can copy dynamically the CDF , postload, and any files needed by postload into /var/tmp during the installation process. The preinstall file itself cannot be invoked from /var/tmp as it is the only mechanism available to move files into /var/tmp.

This feature is valuable for users repackaging the Digital UNIX operating system and who are providing the CDF and user-supplied files on the CD-ROM. When there is a need to modify or select a CDF or postload file as part of the installation process, a writable location is needed because the CD-ROM cannot be written to. For example, assume that several CDFs are shipped on the CD-ROM for the purpose of supporting different hardware or configurations from one distribution media. In this case, you can create a preinstall file that examines the system on which the installation is being executed, and based on the examination, select the appropriate CDF file from among those shipped. The preinstall file can then copy this CDF to /var/tmp/install.cdf where it will later be read by the installation process. Similarly, the preinstall file could choose from among several postload files and copy the one you want to /var/tmp/postload.

The preinstall script should assure that files copied to /var/tmp have the appropriate permission codes (chmod 777 * is the safest way to ensure appropriate permissions).


[Return to Library]  [TOC]  [PREV]  --SECT  SECT--  [NEXT]  [INDEX] [Help]

C.11.4    Burning the CDF and Files on to a CD-ROM

You can repackage the Digital UNIX Version 4.0B operating system CD-ROM to include CDFs and user-supplied files in the /isl directory.

Note

Copying software may be done only for the purpose of licensed use of Digital UNIX. A valid license agreement must be present for all instances of use of the copied Digital UNIX operating system.


Use the method you usually use to burn a CD-ROM (i.e., write to a CD-ROM) if you plan to provide the install.cdf, preinstall, and postload files on a CD-ROM. The method you use depends upon the type of CD-ROM burner you have.

The basic steps to create an image and burn a CD-ROM are:

  1. Mount the Digital UNIX Version 4.0B CD-ROM to determine how much disk space is required on the magnetic disk to which you will be copying the contents of the CD-ROM. For example, to mount the CD-ROM in drive /dev/rz4c on the directory /mnt, enter commands similar to the following:
    # mkdir /mnt
    
    # mount -r /dev/rz4c /mnt
    
    # cd /mnt

  2. Enter the following command to determine disk space in kilobytes:
    # df -k
    Remember this figure and make sure you have a disk large enough to meet the space requirement.

  3. Unmount the CD-ROM using commands similar to the following:
    # umount /mnt

  4. Create an image of the operating system by copying the contents of a Digital UNIX Version 4.0B CD-ROM on to a disk that is at least as large as the figure obtained in Step 2. Use commands similar to the following to copy the contents of the CD-ROM to disk. In the example, the input file is the CD-ROM device, (/dev/rz4c), the output file is the magnetic disk (/dev/rz2c), and the input and output block size is 32 kilobytes (32k).
    # dd if=/dev/rz4c of=/dev/rz2c bs=32k

    Caution

    The output file (of=) must specify a disk partition that starts at block zero (usually a or c). Specifying a partition that does not start at zero (0) results in an operating system image that is not bootable.


  5. Mount the disk to which you just copied the contents of the Digital UNIX Version 4.0B CD-ROM, and use the cp command to copy the install.cdf, preinstall, postload files and any files called by the files into the /isl directory of the image:
    # mount /dev/rz2c /mnt
    
    # cp ./preinstall /mnt/isl/preinstall
    
    # cp ./install.cdf /mnt/isl/install.cdf
    
    # cp ./postload /mnt/isl/postload
    
    # cp ./file_name /mnt/isl/file_name

  6. Depending upon the type of CD-ROM burner you have, use the recommended method to burn a CD-ROM from the modified image on the disk.

    Note

    To ensure that you have a valid, bootable Digital UNIX Version 4.0B image, Digital recommends that you verify the ability to boot from the image on the disk before burning the CD-ROM.