The Certified Systems Administrator certification exam from Hewlett-Packard is designed for UNIX professionals. This exam is based on HP-UX 11iv1 and 11iv2. There is only one exam to pass to get this certification. There are total 78 questions on the exam. Question formats include multiple choice, drag-and-drop, and graphical. A minimum score of 70% is required to pass the exam within 105 minutes. The official exam objectives are given below and are covered in various chapters throughout the book.

1.0 HP-UX System Administration

1. Architecture and structure of HP-UX
2. Describe the HP-UX OS architecture
3. Describe the HP-UX processor families
4. Describe the major hardware components found in HP's current systems
5. Describe the features and benefits of HP disk management solutions
6. Describe the significance of basic LVM concepts and structure
7. Describe the significance of basic VxVM concepts and structure

 2.0 HP-UX user environment, basic commands, and utilities

1. Login and logout of an HP-UX system
2. Determine basic information about a system
3. Execute HP-UX commands from the command line
4. Manage and manipulate files and directories
5. Define and describe the attributes of basic system components
6. Demonstrate the tools and techniques used to identify, monitor, and terminate programs and processes
7. Identify and explain how and when to use advanced shell features
8. Describe and demonstrate how to communicate with system users
9. Describe when and how to access basic network services

 3.0 HP-UX system administration and operational tasks

1. Boot, reboot, and shutdown an HP-UX system or partition
2. Connect and configure HP-UX hardware
3. Describe, configure, and manage HP-UX device files
4. Configure and manage disks and partitions
5. Maintain file and file system integrity and design
6. Backup and recover data on an HP-UX system
7. Create and manage swap space
8. Create and manage user/group environments
9. Describe print spooling and its use
10. Configure and reconfigure the HP-UX kernel
11. Describe common areas of performance bottlenecks
12. Describe SYS-V IPC services and their use
13. Describe SAM and its use
14. Describe System Startup model and its use
15. Describe chroot and its use
16. Describe /etc/default and its use
17. Monitor system activity and events
18. Implement HP partitioning solutions

 4.0 HP-UX system network administration tasks

1. Describe MAC addressing and its use
2. Describe IP addressing and its use
3. Describe NIC and its use
4. Enable DHCP for NIC address configuration
5. Check local connectivity to a known neighbor by IP or by MAC address
6. Describe set_parms and its use
7. Describe common network configuration files and their use
8. Describe network monitor utilities and their use
9. Configure and monitor network services
10. Describe the HP CIFS product suite and its use
11. Describe the ONC suite of network services and their use
12. Describe LDAP and its use
13. Describe DNS and its use
14. Describe sendmail and its use as a MTA and MDA
15. Describe NTP and its use

 5.0 HP-UX installation, upgrade, and recovery tasks

1. Perform an HP-UX installation from local installation media
2. Perform HP-UX software installation from an SD-UX server
3. Configure a SD-UX depot server
4. Install HP-UX patches
5. Describe the features and benefits of Ignite-UX
6. Perform an HP-UX installation from an Ignite-UX Server
7. Install and configure an Ignite-UX server

 6.0 HP-UX security administration tasks

1. Cite 'users level' security settings
2. Describe trusted system and its use
3. Describe password creation options and their use
4. Describe how to implement system access restrictions
5. Describe common system security concerns
6. Describe ssh and its use
7. Describe PAM and its use
8. Describe available security tools
9. Describe common administrative security tasks
10. Explain how various network architectures/features can affect a system security policy
11. Describe Standard Mode Security Extensions

 7.0 Describe HP-UX High Availability and Clustering Features

1. Explain Key HA Terms
2. Identify the Risks with SPOF

 Visit www.hp.com/go/certification for up-to-date and more in-depth information about the CSA exam requirements.

Exam Fee and How to Register for the Exam

The exam fee is US$124. To register for the exam, call Prometric at 1 800 718-3926 or visit http://www.2test.com/ or www.prometric.com. A list of Authorized Prometric Testing Centers (APTCs) can be obtained from these sites as well. You are required to provide your HP student ID when you register for the exam.

HP Student ID

Before registering for the HP CSA HP0-095 certification exam, you must obtain an HP Student ID. This unique identification number is allocated to each HP-UX exam candidate. Visit www.hp.com/certification/americas/student_id.html and follow the procedure.

Additional Exam Objectives for HP0-A01

1.0 HP-UX Architecture and Structure (17%)

1. Describe virtualization technologies and concepts
2. Describe the concept and benefits of the Mass Storage Stack
3. Describe 11i v3 performance and capacity improvements and workload benefits

3.0 HP-UX system administration and operational tasks (26%)

1. Describe Web-Based Enterprise Management (WBEM) and its use
2. Describe HP Systems Insight Manager (HP SIM) and its use
3. Describe System Management Homepage (SMH) and its use

6.0 HP-UX security administration tasks (11%)

1. Use Bastille for system hardening
2. Use Encrypted Volume and File System (EVFS)
3. Identify Identity Management features and functions
4. Use Install-time Security
5. Identify new security features (new to 11iv3)
6. Use the Software Assistant (SWA)
7. Describe Local User Administration

8.0 HP-UX Partitioning Continuum

This chapter covers the following major topics:

- Understanding HP-UX partitioning continuum

- Benefits of partitioning

- Introduction to Management Processor

- Node partitioning - concepts, creating, modifying, and removing

- Virtual partitioning - concepts, creating, modifying, and removing

- Virtual Machines partitioning - concepts

- Resource partitioning - concepts

8.1 Understanding HP-UX Partitioning Continuum

HP-UX 11i Operating Environment software offers Partitioning

Continuum technologies. These technologies enable you to logically

divide a server complex into several smaller computers, with each

allotted dedicated (or shared) hardware resources and run an

independent HP-UX 11i OE instance. Each of these virtual computers

can be used to run a unique application completely independent of

applications running in other virtual computers.

8.1.1 Supported Partitioning Continuum Technologies

Currently, the following four partitioning techniques are available:

1. Node Partitioning - used on both HP Integrity and HP 9000 cell-based servers.

2. Virtual Partitioning - used on both HP Integrity and HP 9000 mid-range and high-end servers.

3. Virtual Machines Partitioning - used on HP Integrity servers only.

4.Resource Partitioning - used on both HP Integrity and HP 9000 servers via HP Process Resource Manager (PRM) and Work Load Manager (WLM) software.

8.1.2 Benefits of Partitioning

Partitioning continuum technologies offer several benefits, such as:

- Better system hardware resource utilization.

- Online allocation, deallocation, and administration of hardware resources.

- Separate, independent OE instance in each partition.

- Different patch level in each partition.

- Dissimilar kernel in each partition.

- Unique application in each partition with complete isolation from other applications.

- Software fault isolation.

Note that not all partitioning techniques offer all these benefits. Most are associated mainly with the first three. The fourth resource partitioning - supports only a subset.

This chapter covers the first two partitioning techniques - node and virtual partitioning - at length; the rest are briefly touched upon.

Before getting into the details of partitioning, let us study something called Management Processor.

8.2 Introduction to Management Processor (MP)

The Management Processor(MP) is part of all new HP Integrity and 9000 series servers. It is a small hardware module that is installed in servers for system management purposes. It has its own processor and can be rebooted independent of the server in which it is installed without impacting the server operation. Through MP, you can access a server complex and perform tasks, such as:

- Access server or node partition (nPar) console.

- Access virtual partition (vPar) console.

- Power off, reset, and send Transfer Of Control (TOC) signal to the system or nPar.

- Send event notification.

- View and log console messages, system events, and error messages.

- View updated cell board and nPar status.

- View cell configuration information.

- View updated boot information through Virtual Front Panel (VFP).

- Display environment parameters, such as electrical power and cooling.

- View and modify LAN MP TCP/IP settings for over the network telnet access.

On older server models, MP is called Guardian Service Processor (GSP). MPcan beaccessed via serial or Ethernet port. Both ports are part of the MP module and dedicated for MP access. For local MP access, you can physically connect any serial display device to the serial MP port. On some server models, this port is referred to as Console/Remote/UPS. For remote, over the network MP access, you need to configure TCP/IP parameters on the LAN MP port. You also need to ensure that proper network connectivity is in place to access the MP remotely using the configured IP address.

At this point, it is important to highlight the differences between LAN MP port and other network ports available in the server or nPar. Although both require configuring TCP/IP parameters, the LAN MP port is dedicated for MP operations and console access. It is available even if the server is down. In contrast, other network ports provide access into the server or nPar and only when the server or nPar is up and in functional state. Users and applications employ these network ports.

8.2.1 Setting Serial MP Display Parameters

It is important to set correct parameter values on the serial console device to allow text and menus to be displayed correctly and navigated properly.

Follow the steps below to perform key console settings on an HP dumb terminal device.

- Connect the HP dumb console terminal to the serial MP/console port on the system.

- Power it on.

- Press User / System key located in the middle of the top row on the console keyboard.

- Press F4 to select Remote Mode. Make sure that the asterisk ( * ) character appears beside it.

- Press F8 to go to Config Keys and then F3 for Datacomm Config. Set BaudRate to "9600", Parity/DataBits to "None/8", RecvPace to "Xon/Xoff", and XmitPace to "None". Press F1 to save when done.

- Press F8 again to go to Config Keys and then F5 for Terminal Config. Set Datacomm/ExtDev to "Serial (1) / Serial (2)", Terminal Id "70096", Keyboard "USASCII", Language "ENGLISH", LocalEcho "off", and CapsLock to "off".

This completes the setup of the display terminal and now it should work well.

8.2.2 Setting LAN MP Parameters

If you wish to access MP over the network as well, you need to connect a display terminal device to the serial MP port and log in to MP (follow procedure in the next sub-section on how to log in to MP). Go to the command mode by typing cm at the MP prompt (refer to Table 8-1 on various commands and options available at the MP level). Execute the lccommand to set appropriate network parameters, such as IP address, subnet mask, default gateway, and MP hostname. Execute the xd command when done to reset and restart MP.

This completes the setup of LAN MP port. You should now be able to access the server or nPar remotely from over the network via this interface.

8.2.3 Interacting with MP

MP is accessed via user accounts. Up to 32 user accounts can be defined, with each user may be set to have different level of access to server complex and MP commands. There exists one powerful pre-configured user called Admin with full access to entire server complex. Other user accounts that you may wish to create can have operator or nPar-specific access.

MP user accounts have nothing to do with HP-UX user accounts.

To access MP via serial port, type in "Admin" as username and "Admin" as password at the MP login prompt. The MP> main menu will appear.

For over the network access, supply the LAN MP IP address with the telnet command to get the MP login prompt. Enter the username and password, as above, to get the MP> prompt.

The following sample is taken from an rp8400 server:

MP login: Admin

MP password: *****

Welcome to the

Rp8400 Management Processor

(c) Copyright 1995-2002 Hewlett-Packard Co., All Rights Reserved.

Version 4.20

MP MAIN MENU:

CO: Consoles

VFP: Virtual Front Panel (partition status)

CM: Command Menu

CL: Console Logs

SL: Show chassis Logs

HE: Help

X: Exit Connection

MP:>

If you are already connected and logged in to the MP, you need to press ctrl+b to get the MP prompt

The main menu options are CO, VFP, CM, CL, SL, HE, and X. Table 8-1 describes them.

Table 8-1 MP Commands

At times, it is necessary to press ctrl+E+c+f to gain access to the console.

8.3 Hardware Available to Work with Partitioning

In the next two sections, node and virtual partitioning techniques are discussed. In order to explain various administrative operations, example commands are executed on an RP8400 server and output of some of them displayed here. Note that most tasks can be performed on Integrity machines the way they are performed on 9000 systems. The RP8400 server has the following hardware configuration:

Hardware model: HP 9000 RP8400 Number of cabinets: 1

Number of server expansion units: 1 Number of cell boards: 4

Number of processors: 16 (4 on each cell board) Memory on each cell board: 8GB

Total amount of memory: 32GB Number of I/O chassis: 4

Number of I/O slots: 32 (8 in each I/O chassis) Maximum nPars this server supports: 4

Maximum vPars this server supports: 16

Let us now examine the two more commonly employed partitioning techniques - node partitioning and virtual partitioning - under HP-UX Operating Environment software.

8.4 Node Partitioning

Node partitioningisa hardware partitioning technique used to divide a cell-based mid-range or high-end server into several smaller computers called node partitions (nPars). Each nPar can then be used as a separate, independent, standalone server running HP-UX OE instance with its own dedicated processors, memory, and I/O chassis. Depending on the hardware model, a server complexmay be divided into as many electrically-separated nPars as the total number of cell boards in the complex. The minimum number of nPars that can be created in a server complex is 2 and the maximum in the largest server complex is 16. Each nPar must have (at least) one core I/O card installed in slot 0 of its I/O chassis. Figure 8-1 shows a logical nPar running two applications.

Figure 8-1 nPar Logical

In the following sub-sections, nPar administration tasks, such as creating the genesis partition, creating an nPar, changing an nPar name, adding a cell to an nPar, removing a cell from an nPar, and removing an nPar, are explained. You can issue commands to perform these operations from any of the nPars within the server complex. Alternatively, you can use the GUI software called Partition Manager to perform these tasks.

Some of the nPar administration commands are given in Table 8-2.

Table 8-2 nPar Administration Commands

8.4.1 Creating the Genesis Partition

The first nPar created in a server complex is always referred to as the Genesis Partition. Itis a single-cell partition built via MP. Once created, you can load the HP-UX 11i OE into it.

To create the genesis partition, go to MP --> CM --> CC --> and choose option G. Pick a cell that contains processors and memory, and must be connected to an I/O chassis with a core I/O card installed. When done, use the bo MP command to boot the nPar. Insert the HP-UX 11i DVD installation media in the drive and follow the procedure outlined in Chapter 09 "HP-UX Operating Environment Installation" to install the OE in the nPar.

After the installation is over, log in to the genesis partition, and issue the parstatuscommand to check the status and resources available in it.

# parstatus -w

The local partition number is 0.

# parstatus -Vp0

[Partition]

Partition Number : 0

Partition Name : Partition 0

Status : active

IP address : 0.0.0.0

Primary Boot Path : 0/0/0/2/0.6.0

Alternate Boot Path : 0/0/0/3/0.6.0

HA Alternate Boot Path : 0/0/0/0/0.0.0

PDC Revision : 17.8

IODCH Version : 5E40

CPU Speed  : 750 MHz

Core Cell : cab0,cell0

[Cell]

 CPU Memory Use

 OK/ (GB) Core On

Hardware Actual Deconf/ OK/ Cell Next Par

Location Usage Max Deconf Connected To Capable Boot Num

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

cab0,cell0 active core 4/0/4 8.0/ 0.0 cab0,bay0,chassis0 yes yes 0

[Chassis]

 Core Connected Par

Hardware Location Usage IO To Num

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

cab0,bay0,chassis0 active yes cab0,cell0 0

By default, the genesis partition is called “Partition 0”.

As you can see from the output that there are 4 processors in the

genesis partition and 8GB of memory. It also shows the primary and

alternate boot disk paths.

8.4.2 Creating an nPar

Once the genesis partition is formed, you can make additional nPars using unused cell boards and I/O chassis in the server complex.

Let us create a single cell nPar called test_npar. Log in to the genesis partition and perform the following steps:

1. Identify what cell board is available. Run the parstatus command:

# partstatus -AC

Cell 1 is available.

Cell 2 is available.

Cell 3 is available.

2. Determine which I/O chassis is free. Do the following:

# parstatus -AI

I/O chassis 1 is available.

I/O chassis 2 is available.

I/O chassis 3 is available.

3. Execute the parcreate command to create test_npar. Use cell board 1 and I/O chassis 1. Specify nPar name with the -P option and cell number to include with the -c option. Values following each colon character are "base" (for cell type), "y" (to include the cell in next boot), and "ri" (reuse interleave) to reuse memory after a failure). Since these are default values, they are not exhibited in the following command line:

# parcreate -P test_npar -c1:::

Partition Created. The partition number is: 1

4. Define primary (-b option) boot disk using the parmodify command. If you wish to use alternate boot disk as well, define it using the -t option.

# parmodify -p1 -b 1/0/0/2/0.6.0 -t 1/0/0/3/0.6.0

5. Verify that the nPar is created successfully. Use the parstatus command.

# parstatus -Vp1

[Partition]

Partition Number : 1

Partition Name : test_npar

Status : active

IP address : 0.0.0.0

Primary Boot Path : 1/0/0/2/0.6.0

Alternate Boot Path : 1/0/0/3/0.6.0

HA Alternate Boot Path : 0/0/0/0/0.0.0

PDC Revision : 17.8

IODCH Version : 5E40

CPU Speed : 750 MHz

Core Cell : cab0,cell1

[Cell]

 CPU Memory Use

 OK/ (GB) Core On

Hardware Actual Deconf/ OK/ Cell Next Par

Location Usage Max Deconf Connected To Capable Boot Num

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

cab0,cell1 active core 4/0/4 8.0/ 0.0 cab0,bay0,chassis1 yes yes 1

[Chassis]

 Core Connected Par

Hardware Location Usage IO To Num

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

cab0,bay0,chassis1 active yes cab0,cell1 1

This nPar also has 4 processors and 8GB of memory.

# parstatus -P

[Partition]

Par # of # of I/O

Num Status Cells Chassis Core cell Partition Name (first 30 chars)

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

0 active 1 1 cab0,cell0 Partition 0

1 active 1 1 cab0,cell1 test_npar

You need to install the HP-UX 11i OE in the new nPar now. Use the bo MP command to boot the nPar. Make sure that you have the HP-UX 11i DVD installation media in the drive. Follow the HP-UX 11i OE installation procedure outlined in Chapter 09 “HP-UX Operating Environment Installation” to install the OE in the nPar.

8.4.3 Changing an nPar Name

To change the first (genesis) nPar namefrom Partition 0 to prod_npar, usethe parmodify command. Verify results with parstatus after the operation is complete.

# parmodify -p0 -P prod_npar

# parstatus -P

[Partition]

Par # of # of I/O

Num Status Cells Chassis Core cell Partition Name (first 30 chars)

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

0 active 1 1 cab0,cell0 prod_npar

1 active 1 1 cab0,cell1 test_npar

8.4.4 Adding a Cell to an nPar

To add a cell to an active nPar (say test_npar), use the parmodify command. Specify partition number with -p option and cell number with -a option. Make sure that the cell (cell 2 in this case) is available. The –B option activates the nPar after it is rebooted. You must reboot the nPar for reconfig (option -R with the shutdown command) before the modification takes place.

# parmodify -p1 -a2::y: -B

In order to activate any cell that has been newly added, reboot the partition with the -R option.

Command succeeded.

# shutdown -Ry now

When the nPar is back up, execute parstatus and verify added hardware.

# parstatus -P

[Partition]

Par # of # of I/O

Num Status Cells Chassis Core cell Partition Name (first 30 chars)

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

0 active 1 1 cab0,cell0 prod_npar

1 active 2 2 cab0,cell1 test_npar

To add a cell to an inactive nPar, you have two options. If you wish to boot up the nPar and activate it after the modification is done, use the parmodify command as above. Do not specify –B option if you do not wish to boot and activate the nPar.

8.4.5 Removing a Cell from an nPar

To remove an active cell from an activenPar, use -d option with parmodify and specify the cell number (cell 2 in this case). You must reboot the nPar for reconfig (-R option with shutdown) for the changes to take effect.

# parmodify -p1 -d2 -B

command succeded

# shutdown -Ry now

To remove an inactive cell from an active or inactive nPar, use -d option and specify the cell number to delete. No need to specify -B option and no need to reboot.

# parmodify -p1 -d2

8.4.6 Removing an nPar

To remove an active nPar, run parremove from the nPar you wish to delete and supply -F option. This operation unassigns all cells from the nPar and destroys its partition definition. You must then shutdown and halt the nParfor the changes to take effect.

# parremove -Fp1

Use "shutdown -R -H" to shutdown the partition.

The partition deletion will be effective only after the shutdown.

# shutdown -RH now

To remove an inactive nPar, run parremove, as above, from an active nPar and specify the partition number of the nPar you wish to delete. No need to use –F option and no need to shutdown the inactive/deleted partition since it is already down.

8.4.7 Using Partition Manager GUI

The Partition Managersoftwareis loaded as part of the HP-UX 11i OE installation on supported servers. You can run it either in graphical or textual mode via SAM as follows:

Go to SAM --> Partition Manager.

You can run it in graphical mode using a browser or parmgr command as well. The browser interface requires that you have Apache web server setup, and up and running.

To start partition manager using the browser interface, set the variable WEBADMIN to 1 in the /etc/rc.config.d/webadminfile and execute associated startup script.

# vi /etc/rc.config.d/webadmin

WEBADMIN=1

# /sbin/init.d/webadmin start

/usr/obam/server/bin/apachectl start: httpd started

The default port number is 1188. If the system or nPar name is hp10, for instance, you would type the following URL in the browser to run it:

http://hp10:1188/parmgr

The parmgr command enables you to run partition manager fromthe command prompt. Figure 8-2 exhibits partition manager GUI.

# /opt/parmgr/bin/parmgr

Figure 8-2 Node Partition Manager GUI

From the two graphical interfaces you can create, view, modify, and remove nPars, among other nPar administrative activities.

8.5 Virtual Partitioning

Virtual partitioningisa software partitioning technique used to divide a server or an nPar into several virtual, smaller computers called Virtual Partitions (vPars). Each vPar can then be used as a separate, independent, standalone server running HP-UX OE instance with its own processor(s), memory, and I/O slot(s). The number of vPars that can be created in a server or an nPar is equal to the number of processors you have in the server or nPar. In other words, each vPar requires at least one processor to be created. The minimum number of vPars that can be defined in a server complex is 2 and the maximum in the largest server complex is 128. Figure 8-3 exhibits a server/nPar running two applications. This server/nPar is divided into two vPars, each with different patch level and running different applications.

Figure 8-3 vPar Logical

When vPars are created, a software piece calledvpmonis enabled, which sits between the server firmware and HP-UX OE instances running in vPars. When a vPar boots up, vpmon looks into the vPar database, /stand/vpdb, and loads the OE in the vPar. It uses the boot disks defined for the vPar and assigns the hardware components configured for it.

In the following sub-sections, vPar administration tasks, such as creating a vPar, adding a processor to a vPar, removing a processor from a vPar, adding an I/O path to a vPar, removing an I/O path from a vPar, adding memory to a vPar, resetting a vPar, and removing a vPar, are explained. You can issue commands to carry out these operations from any of the vPars within the server complex. Alternatively, you can use the GUI software called Virtual Partition Manager to accomplish these tasks.

Some of the more common vPar administration commands you are going to be interacting with are given in Table 8-3.

Table 8-3 vPar Administration Commands

You can create vPars on either a supported server (not already node-partitioned) or in an nPar. Make sure that you have VPARMGR and T1335AC software installed. Issue the swlist command as follows to check whether or not the two software components are loaded:

# swlist | grep VPARMGR

VPARMGR B.11.11.01.02 Virtual Partition Manager - HP-UX

# swlist | grep T1335AC

T1335AC A.02.02.00 HP-UX Virtual Partitions

Install them using the swinstall command if they are not already loaded. Consult Chapter 10 “Software Management” on how to install software.

8.5.1 Bound and Unbound Processors

When working with vPars, youcanassign processors as either bound or unbound.

A processor assigned to a vPar to handle both normal processing as well as I/O interrupts is referred to as a bound processor. Every vPar must be assigned at least one bound processor. A processor should be allocated as a bound processor to a vPar if the vPar runs applications which require both processor and I/O horsepower. A processor can be added online to a vPar as a bound processor, however, a bound processor cannot be removed online from a vPar.

An unbound processor, on the contrary, is either not assigned to a vPar, or assigned, but does not handle the vPar’s I/O interrupts. A processor should be allocated as an unbound processor to a vPar if the application running in the vPar is processor-intensive but not I/O-intensive. An unbound processor can be added to or removed from a vPar online.

8.5.2 Creating the First vPar

To understand vPar operations, let us assume that an nPar, containing 2 cell boards with 8 processors, 16GB of memory, and 2 I/O chassis, is available in the RP8400. The hardware addresses of the processors are 41, 45, 101, 109, 141, 145, 201, and 209. Use the vparcreate commandto create a vPar called test_vpar0 with 1 processor bound to it and 1 unbound. Set the minimum and maximum limit on the number of processors this vpar can get to 1 and 2, respectively. Allocate 8GB of memory and define hardware I/O paths 1/0/0/2 and 1/0/0/3. Use 1/0/0/2/0.6.0 as primary boot disk. With –B option, specify that the vPar, at each reboot, is going to search automatically for the bootable devices and boot from the one it finds first. When the following is executed, all the configuration information gets stored in the vPar databaselocated in /stand/vpdb file. If this file does not exist, it will be created.

# vparcreate -p test_vpar0 -a cpu::2 -a cpu:::1:2 -a cpu:41 -a mem::8192 -a 

io:1/0/0/2 -a io:1/0/0/3 -a io:1/0/0/2/0.6.0:BOOT -B search -B auto

The options used with vparcreate are explained below:

-p - vPar name.

-a cpu::2 - number of processors to be allocated.

-a cpu:::1:2 - minimum and maximum limits on processor allocation.

-a cpu:41 - hardware address of the processor to be bound. If not explicitly defined, the processor will be assigned as unbound.

-a mem::8192 - amount of memory, in MBs, to be allocated.

-a 1/0/0/2 - all devices physically located on the path will become available to this vPar.

-a 1/0/0/2/0.6.0:BOOT - primary boot path.

-B search - automatic search for the boot device at boot time is enabled.

-B auto - autoboot is enabled.

The last three functions can also be performed using the setboot command, the vparmodify command,

or commands at the BCH or EFI menu. Detailed discussion on BCH and EFI is covered in Chapter 17

“Shutting Down and Starting Up an HP-UX System”.

Modify the AUTOfile on the primary boot path (/dev/rdsk/c0t6d0 in our case) on the server or nPar in which you created the vPar. Use the mkboot command. This is done to make the system aware that vPars are implemented and make sure that the system loads vpmon first at next system reboot before it starts loading the HP-UX kernel.

On HP 9000 servers, do the following:

# mkboot -a "hpux /stand/vpmon -a" /dev/rdsk/c0t6d0

On HP Integrify servers, do the following:

# mkboot -a "boot vpmon -a" /dev/rdsk/c0t6d0

This creates a vPar within the nPar with 2 out of 8 processors and 8GB out of 16GB memory

allocated to it. Now each time you reboot this nPar, it will come up as a vPar.

To manually boot test_vpar0, go to the nPar console,press ctrl+a to get the Virtual

Console monitor MON> prompt, and run the vparload command as follows:

MON> vparload -p test_vpar0

Alternatively, you can boot the vPar from ISL prompt (in case of HP 9000 server):

ISL> hpux /stand/vpmon vparload -p test_vpar0

Once test_vpar0 is up and functional, run vparstatus to view status information.

# vparstatus

[Virtual Partition]

 Boot

Virtual Partition Name State Attributes Kernel Path Opts

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

test_vpar0 Up Dyn,Auto /stand/vmunix 

[Virtual Partition Resource Summary]

 CPU Num Memory (MB)

 CPU Bound/ IO # Ranges/

Virtual Partition Name Min/Max Unbound devs Total MB Total MB

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

test_vpar0 1/ 2 1 1 8 0/ 0 8192

Under Attributes column, two values – Dyn and Auto – are shown. "Dyn" means dynamic hardware configuration changes are allowed for the vPar and "Auto" indicates that the vPar is auto-bootable.

Run vparstatus againto display the information in a different way.

# vparstatus -vp test_vpar0

[Virtual Partition Details]

Name: test_vpar0

State: Up

Attributes: Dynamic,Autoboot

Kernel Path: /stand/vmunix

Boot Opts:

[CPU Details]

Min/Max: 1/2

Bound by User [Path]:

Bound by Monitor [Path]: 41

Unbound [Path]: 45

[IO Details]

 1.0.0.2

 1.0.0.3

 1.0.0.2.0.6.0 BOOT

 [Memory Details]

Specified [Base /Range]:

 (bytes) (MB)

Total Memory (MB): 8192

8.5.3 Creating Another vPar

At this point, you may wish to createanother vPar using some or all available resources. Execute vparcreate command as follows from test_vpar0 to create test_vpar1:

# vparcreate -p test_vpar1 -a cpu::2 -a cpu:::1:2 -a cpu:101 -a mem::8192

-a  io:2.0.0.2 -a io:2.0.0.3 -a io:2/0/0/2/0.6.0:BOOT -a io:2/0/0/3/0.6.0:ALTBOOT 

-B search -B auto

This creates test_vpar1 within the nPar. Note that this command sets up primary as well as alternate boot disk paths.

You now need to install HP-UX 11i OE in test_vpar1. Make sure that you include vPar software when you install the OE. Shutdown test_vpar0, go to the BCH prompt, and install HP-UX 11i OE on one of the boot disks assigned to test_vpar1. Use the DVD drive available to the nPar. When installation is complete, go to the virtual monitor prompt and boot test_vpar0.

MON> vparload -p test_vpar0

Log on to test_vpar0 when it is up and execute the mkboot command

to modify the AUTO file contents on test_vpar1’s boot disks.

# mkboot -a "hpux /stand/vpmon -a" /dev/rdsk/c2t6d0

# mkboot -a "hpux /stand/vpmon -a" /dev/rdsk/c3t6d0

Now run vparboot from test_vpar0 to bring test_vpar1 up.

# vparboot -p test_vpar1

When test_vpar1 is up, use vparstatus to verify the configuration and hardware allocation.

8.5.4 Adding a CPU to a vPar

The test_vpar0 vPar currentlyhas 1 bound (hw path 41) and 1 unbound (hw path 45) processors. In order to add a bound processor, you need to first increase the maximum processor allocation limit to at least 3 from current 2, and then add the processor. This operation requires that test_vpar0 is down. Do the following on test_vpar0:

# shutdown -hy now

Log on to test_vpar1 and execute the vparmodify commandas follows. The first command increases the maximum limit to 3, the second adds the processor at hw path 101 as bound, and the third boots up test_vpar0.

# vparmodify -p test_vpar0 -m cpu:::1:3

# vparmodify -p test_vpar0 -a cpu:141

# vparboot -p test_vpar0

When test_vpar0 is back up, run vparstatus with -vp options to verify the modifications.

To add a processor as unbound, you do not need to shutdown test_vpar0. First, increase the processor limit to 4 (since you now have 3 processors assigned) and then add a processor. When the hardware path of a processor is not explicitly defined, vparmodify adds the processor as unbound.

# vparmodify -p test_vpar0 -m cpu:::1:4

Now add a processor. Do either of the following two. Both produces the same result. The first command increases the total count of processors to 4 (from current 3) and the second simply adds one processor.

# vparmodify -p test_vpar0 -m cpu::4

# vparmodify -p test_vpar0 -a cpu::1

8.5.5 Removing a Processor from a vPar

At this point, you have fourprocessors assigned to test_vpar0 (2 bound and 2 unbound). To remove the bound processor (141) from test_vpar0, shut it down.

# shutdown -hy now

Run vparmodify on test_vpar1 and specify the affected vPar name with –p option.

Use –d to furnish the hardware path of the bound processor to be removed.

# vparmodify -p test_vpar0 -d cpu:141

# vparboot -p test_vpar0

To remove an unbound processor, do one of the following on test_vpar0:

# vparmodify -p test_vpar0 -m cpu::2

# vparmodify -p test_vpar0 -d cpu::1

This reduces the total number of processors in test_vpar0 from current 3 to 2. No need to shutdown.

Check the results using vparstatus with -vp options.

8.5.6 Adding an I/O Path to a vPar

To add an I/O path to test_vpar0, shut it down, and execute the following on test_vpar1. When done, boot test_vpar0from test_vpar1. Check the results using vparstatus with –vp options.

# vparmodify -p test_vpar0 -a io:3.0.0.3

# vparboot -p test_vpar0

8.5.7 Removing an I/O Path from a vPar

To remove an allocated I/O path from test_vpar0, shut itdown and execute the following on test_vpar1. When done, boot it up from test_vpar1. Verify the results using vparstatus with –vp options.

# vparmodify -p test_vpar0 -d io:3.0.0.3