LDOMS I/O Best Practices - Data Reliability With Logical Domains 
by Peter A. Wilson
July, 2008
Datacenter best practices for disk and network I/O are relatively well defined. Configure multiple data paths to network and storage resources to increase availability. For data reliability, use media redundancy including hardware and software RAID and sophisticated filesystems such as the Zettabyte File System (ZFS). These rules of thumb drive the way in which real-world datacenter servers are configured and cabled — but how do these rules translate into the virtual world? What does it mean to have redundant virtual disks and networks? How should they be leveraged to provide availability and reliability benefits similar to those in the physical world? This Sun BluePrints™ Series article begins to address the answers to these questions in the context of Logical Domains.
- Data Reliability With Logical Domains
- About This Article
- Data Availability And Reliability Overview
- Internal Storage Reliability With Logical Domains
- Hardware RAID
- Reliability In I/O Domains
- Reliability In Guest Domains
- I/O Reliability And Availability On Sun Platforms
- Example: Sun SPARC Enterprise T2000 Server I/O Architecture
- I/O Architecture Summary
- Installing And Configuring Logical Domains
- Configuring Hardware RAID
- About Hardware RAID Disk Sets
- Implementing Hardware RAID
- Creating A Logical Domain
- Configuring ZFS In The I/O Domain
- Implementing ZFS And Cloning Guest Domain Volumes
- Create A ZFS Pool
- Clone And Unconfigure
- Clone And Run The Unconfigured System
- Save Your Configuration
- Configuring Volume Management In Guest Domains
- Setting Up Volume Management Through Network Install
- Set Up The Guest Domain
- Rules File Setup
- Profile File Setup
- Sysidcfg File Setup
- Add Install Client
- Install
- Housekeeping
- Test
- Conclusion
- About The Author
- Acknowledgments
- References
- Ordering Sun Documents
Peter A. Wilson has more than 16 years of industry experience, 12 of which have been with Sun, serving in a wide variety of hardware, software, systems and product marketing roles. Peter moved from the United Kingdom to the US in 2000 to lead the customer tests of Sun's Netra™ and fault-tolerant servers. Peter is currently a technical marketing engineer responsible for all aspects of Sun's servers with CoolThreads technology. Peter holds a M.Eng (Master of Engineering) degree in Microelectronics and Software Engineering from the University of Newcastle-upon-Tyne, U.K.
The author would like to thank Steve Gaede, an independent technical writer and engineer, for working through the issues and configurations described in this paper and developing this article based on them. Thanks also to Maciej Browarski, Alexandre Chartre, James MacFarlane, and Narayan Venkat for their help in understanding some of the nuances of using Solaris Volume Manager software.
Sign up or Log in to add a comment or watch this page.