Menu

Search



Using Prism Deploy to Streamline Workstation Builds

Pam
Deploying Tasks to Computers

The information in this article applies to:

• Prism Deploy

Summary

A company is always going through the process of rebuilding existing workstations, upgrading existing hardware to a new OS, or rolling out new hardware to users. Wouldn’t it be nice if this process was somewhat automated? If only you could install software and updates and tweak the system with only a few steps rather than having to connect to the web to download updates, insert CDs to install software, find and type in serial numbers, answer prompts, etc.

Good news!! Prism Deploy is the perfect tool for this process!! And the even better news is that using Prism Deploy to build the computers will help keep the software installs standardized.

Install the OS

The first step for building a system is to install the OS. One way to accomplish this is by using cloning software such as Symantec's GHOST or PowerQuest's DriveImage. These are excellent companion products to New Boundary Technologies’ packaging and deployment products. Use cloning software to quickly build a base Windows system. Then use Packages to easily and accurately install applications and customize the environment.

If the OS build is on diverse hardware, then cloning software may not be the best option. Another common way to quickly install the OS is to implement Microsoft’s unattended installation processes. Refer to Microsoft’s web site for more information. Here are links to a couple of the Knowledge Base articles.

Windows 2000: http://support.microsoft.com/default.aspx?scid=kb;EN-US;q216258

Windows XP: http://support.microsoft.com/default.aspx?scid=kb;en-us;Q314459

Build a Library of Packages

Build Prism Packages for all the changes you wish to make to the computers during the build process. For example, build a Package of the applications that may be installed (office suite, email, accounting, virus software, etc.), Windows OS updates (critical/security updates), and any special tweaks you normally do to computers (standardize the Desktop with company colors, install policies to lock-down the OS, etc.).

It is recommended that you build a single Package for each update in order to more easily manage the Packages, for troubleshooting purposes, and so that each computer can get a subset of the Packages installed. For example, a computer destined for the marketing group won’t likely need the accounting software. This also means building each Package on a “clean” computer. Refer to the related tech note Q10016: "Best Practices for Package Creation" for recommendations when building Packages.

Install the Packages

Now that you have the OS installed and a library of Packages built, how do you go about installing the Packages to the computers? Using Prism Deploy, there are two general paths you can take. Read the steps involved with each path and determine the best solution (or combination of the two) for your company.

Solution 1:

If these computers will be on a TCP/IP network, use Prism Deploy’s Console to deploy packages to newly-built computers. Create a Prism Deploy channel that includes Tasks that represent the library of Packages. Next, create groups that represent different configurations for your environment. For example, you could create functional groups such as Accounting or Human Resources, or geographic groups such as San Francisco Office and San Jose Office. Then assign the appropriate Tasks to the groups you created.

After the OS has been installed on a computer, install Prism Deploy’s client on the computer and subscribe the computer to the channel. This can be accomplished by two methods: a Direct Client push from the Console or by building a Subscription file.

The Direct install requires that the computer hosting the channel and the target computer are both NT, 2000, and XP. The host and target computer should also be on the same Domain, unless you have a trust relationship set up between Domains. You must also be managing the channel while logged in as a user with sufficient rights to the target computer, usually a Domain Administrator. From the Prism Deploy Console simply select “Edit | Add Computer” and follow the prompts. This will install the client on the target computer, subscribe the computer to the channel, add the computer name to the channel database, and reboot the target computer.

A subscription file is another method of installing the client. Build a subscription file by selecting "Deployment | Generate Subscription File" from the Prism Deploy Console. Follow the prompts to build an executable and run that executable on the target computers. The subscription file will install the client, subscribe it to the channel, and add the computer to the channel database.

Once the computer has been added to the channel, simply drag & drop the computer to the appropriate group(s) to assign Tasks. For example, you may have created a general group to install Tasks that all computers will be receiving (Desktop configuration, email and virus software, etc.) so a newly added computer will be added to the “Everyone” group. This particular computer will be going to Marge in payroll so you also assign the computer to the “Accounting” group to get the payroll software.

Prism Deploy will install all Tasks, reboot if necessary and you’re ready to roll out the computer.

Additional Ideas when using the Prism Deploy Console to Build Computers:

·         If a computer that was built using this method ever needs to be rebuilt, simply reinstall the OS and reinstall the client and subscribe it to the channel. The computer will already be a member of the channel and will get the same Tasks installed because of group membership.

·         Windows Service Packs and OS updates can contain hardware-specific changes that may complicate the process of building and installing Packages in a heterogeneous environment. One method for rolling out these changes with Prism Deploy is to create a Command Task that launches the native installation process. Microsoft updates often support “unattended” or “quiet” install switches so there is no need to answer prompts when these tasks are running.

·         When setting up Tasks in Prism Deploy, you may want to configure the Tasks to run in Unattended mode so you don’t need to log in as a user before the Tasks are installed. This is a Property of Prism Tasks. When installing Tasks unattended you will also need to define that the Task run as a domain user or set up NullSessionShares on the network share hosting the Packages.

·         Create a second channel that will be used for maintaining these computers when they are out in the field. This channel can be used to push out updates or new software. Subscribe computers to the second channel before rolling them out.

Solution 2:

If computers are not on a network or you don’t have access to the Prism Deploy Console, another method for installing your library of Packages is to build Prism scripts. Scripts are text files containing Prism commands and have the file extension .PTS.

Create a single script to install all Packages or several scripts representing various builds of the computer (see our Accounting example above). Prism command-line options are documented in Appendix A “Scripting Prism Deploy Tasks”. A typical script might look like the following:

/InstallPackage p:\prism\general\hotfix.pwc

/InstallPackage p:\prism\general\email.pwc

/InstallPackage p:\prism\general\virus.pwc

/InstallPackage p:\prism\accounting\payroll.pwc

/Run /Wait p:\prism\security\W2KSP2.EXE /q /u /o

/InstallPackage p:\prism\general\companydesktop.pwc

Before running the script you must first install the Prism Deploy Client and Service to the target computer if it is Windows NT, 2000, or XP. The Service allows Prism to write changes to locked-down files, folders and registry branches no matter the rights of the currently-logged user. To install the Service, you must be logged in as a user with local administrative rights. Then run the /InstallClient command (refer to Chapter 16 of the Prism Deploy User’s Guide for details on this command-line):

ptclient.exe /InstallClient <source> <destination> /client /service

This will install the client and service locally. (Replace the items in brackets with paths to the client source path and destination where you want the client installed on the local computer.) Optionally, you can also install the Prism shell extensions that allow you to double-click on Packages or scripts to install them. The command-line to install the client, service and shell extensions is:

ptclient.exe /InstallClient <source> <destination> /all

<!--[if !supportEmptyParas]--> <!--[endif]-->

Once the service and other optional components are installed, then launch the script. Without the shell extensions, the command-line is:

ptclient.exe /playscript \\server\prism\script.pts

With the shell extensions installed, simply double-click on the .PTS file to launch it and install all Packages. Note that if various Packages require reboots, the reboot is queued until the script is complete.

Additional Ideas when using the Prism Deploy Scripts to Build Computers:

·         Place the Prism client, script, Packages, etc. on a CD-ROM and kick off the client install and script from the CD. Prism supports a %PT_CD% variable representing the computer’s CD drive letter so replace any paths pointing to the CD drive with this variable. Set up the CD to autorun when inserted.

·         Create a network user account whose login script automatically launches the client install and script. After the OS has been installed, simply log in as the “build” user to install the software and updates.

Feel free to contact the New Boundary Technologies Support Center to share your ideas. We appreciate any feedback you may have.

New Boundary Technologies Support Center
support@newboundary.com
612-379-1851


Also In This Category


On a scale of 1-5, please rate the helpfulness of this article


Not Helpful
Very Helpful
Optionally provide your comments to help us improve this article...

Thank you for your feedback!

Add Your Comments
Name:
Email Address:
RadEditor - HTML WYSIWYG Editor. MS Word-like content editing experience thanks to a rich set of formatting tools, dropdowns, dialogs, system modules and built-in spell-check.
RadEditor's components - toolbar, content area, modes and modules
   
Toolbar's wrapper  
Content area wrapper
RadEditor's bottom area: Design, Html and Preview modes, Statistics module and resize handle.
It contains RadEditor's Modes/views (HTML, Design and Preview), Statistics and Resizer
Editor Mode buttonsStatistics moduleEditor resizer
 
 
RadEditor's Modules - special tools used to provide extra information such as Tag Inspector, Real Time HTML Viewer, Tag Properties and other.
   
Verification Code:
Details
Last Modified: 13 Years Ago
Last Modified By: Pam
Type: INFO
Rated 1 star based on 11 votes.
Article has been viewed 4K times.
Options
Customer Support Software By InstantKB 2015-2
Execution: 0.000. 9 queries. Compression Disabled.