Deployment Framework Best Practices Red Hat Enterprise Linux on HP ProLiant Technical white paper Table of contents Introduction ......................................................................................................................................... 2 Overview ........................................................................................................................................ 2 Intended audience..............................................................................
Introduction Overview This document presents a technical survey of Linux deployment issue, best practices and tools. It includes recommendations for assessing the various requirements of a complete deployment framework, both from a strategic and tactical perspective. Through examination of background concepts and best practices, an example implementation of integrating these diverse offerings into a cohesive deployment framework is presented.
Figure 1: Deployment continuums As one criterion, an environment needs to assess both the current and planned mix of operating systems, and decide if the tendency is towards one or multiple platforms. If you determine that a single operating platform is your desired end state, this commonly yields a smaller set of tools and solutions to consider.
The scale and the longevity of the business service being provided are important in the development of a deployment framework. For these criteria, there are often three differing strategic approaches for deployment processes: • Divergence - for environments comprised of quick cycles and short life times of the resulting service.
Figure 2: Deployment component scope Hardware - The foundation is the hardware platform. It can be either a physical system, a virtualization hypervisor, or even a virtual machine. This is the base element that any deployment framework must interact with. In other words, the component that you unpack from the shipping container that must either have remote, perhaps via network or baseboard management controller ( BMC), or human hands-on access.
You can access the iLO control through any number of means including HP Online Configuration utility, hponcfg, the System Management Architecture for Server Hardware (SMASH) command line protocol, and the Intelligent Platform Management Interface (IPMI). New for the HP ProLiant Gen8 platform, is the embedded HP Intelligent Provisioning ( IP) environment.
specific kernel errata are required or HP and Red Hat codevelop Driver Update Packages (DUP) that deliver updated device drivers necessary to enable the respective HP ProLiant generation server. BEST PRACTICE: HP publishes minimum operating system versions for each ProLiant generation platform, noting any special requirement, such as specific kernel errata or driver updates. You must review any applicable HP Customer Advisory (CA) reports to understand possible required additional actions.
the backup and recovery process. Example implementation With the background concepts previously covered, you should now have a good basic understanding of what a comprehensive deployment framework must provide. All that remains is a practical implementation, providing coverage of these aspects. Given the focus of this document, deployment of HP ProLiant with RHEL, one approach is to utilize as many solution-based offerings from the vendors as possible.
Figure 3: Deployment framework solution blocks Based upon these native building blocks already present, it is fairly straightforward to implement a robust, flexible, and scalable deployment framework. Further, each of these can be extended through the inherent modularity and interfaces, to flesh out the remaining aspects. NOTE: At this point, much of the Red Hat Network Satellite server's software stack is really FOSS, courtesy of Red Hat principles and practices.
Network infrastructure services Starting with the foundational element, the network infrastructure services, you must attend to these very basic services. Even if the network infrastructure services are already provided and no additional work on your part is required, you should understand them as they offer great impact to the overall deployment strategy.
Domain Name Service (DNS) - In order to associate memorable hostnames with IP addresses, DNS provides a way for humans and systems to perform this forward and reverse translation. BEST PRACTICE: It is not uncommon for DNS and DHCP to be more closely coupled, allowing a client system to suggest a hostname and allowing DNS and DHCP to provide an IP address and the associated mapping function. This is known as Dynamic DNS (DDNS).
Figure 5: Deployment framework solution – provisioning The Cobbler toolset provides a way to collect installation media into network repositories, offering standardized ways to present these to target systems. Many different Linux distributions are supported. BEST PRACTICES: 12 • Relying upon both DHCP and PXE from the network infrastructure services, this toolset allows systems to easily perform a network boot and installation process.
With both Cobbler and kickstart in place, the RHEL installer, anaconda, can now perform the install completely without intervention. The anaconda toolset thus provides a scripted installation paradigm, to allow hardware abstraction, yet delivering a correct by construction result.
BEST PRACTICES: • You must develop and adopt a security policy which includes proactive patch management strategies. Once these are in place, the use of the configuration management functionality simply becomes an instrument of implementation. • Due to the rich capabilities of the previously describe provisioning block, you can make the bootstrap process an integral part of the profile, simply embed the steps into the post section of a kickstart file.
Backup / Recovery - Conspicuous by its absence, you might wonder how to address this discipline given the deployment framework solution provided so far. But some of the core attributes are built-in to the solution. Usage of this implementation yields an infrastructure that can easily redeploy a service hosting platform to a known state. This can be a faster and more prudent approach than trying to capture all the entropy of the core platform over time through traditional backup and restore processes.
Figure 7: Deployment framework - implementation In respect to the background material and concept, this implemention definitely focuses onthe Linux operating platform but is clearly mid-range with regard to preferences for solution-oriented versus toolbased. While using vendor supplied offerings, the underpinning technology is mostly FOSS toolsets. Depending upon the configuration management paradigms used in the portion, you can cater to either divergence or convergence disciplines.
Summary The first part of this document gives you a relevant background to map the current landscape of possibilities and the tendencies of the organization. It also gives you an overall representation of the aspects needed by a deployment framework. These are provided to perform a self-assessment and survey approach. The second part of this document presents an example architecture using the background information to provide such an implementation.
For more information For additional information, refer to the resources listed below. Resource description Web address HP ProLiant Gen8 www.hp.com/go/proliantgen8 Linux on HP ProLiant servers http://www.hp.com/go/proliantlinux HP Linux documentation www.hp.com/go/linux-docs, select HP Linux Server Management Software. FOSS http://en.wikipedia.org/wiki/Free_and_open_source_software BMC http://en.wikipedia.org/wiki/Baseboard_management_controller#Baseboard_ management_controller iLO http://www.
Share with colleagues © Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. Trademark acknowledgments, if needed.
