Managing RAID on Linux - Derek Vadala [104]
Choose Configure and then New Configuration from the main menu. A list of disks, organized by channel, is presented. Use the arrow keys to move between disks and the SPACEBAR to include a disk in the array. Select your original system disk first and then the new disk. Once a disk is selected, it will be marked online.
Press the ENTER key twice, and a window containing RAID properties appears. Make certain that RAID-1 is selected, and tune the other options to you meet your needs.
Choose Accept (you will need to press the ENTER key twice) and a final confirmation dialog appears. Answer yes, and the configuration is written into the controller's memory.
At the main menu, choose Check Consistency. Use the SPACEBAR to select the array you just created and press F10 to begin the consistency check. The process could take some time, but when it is complete, your RAID-1 will be operational. Then reboot the system and run Linux as you would normally.
Managing Arrays
Like many other hardware vendors, LSI Logic does not provide a way to manage arrays from within Linux. Dell has ported the MS-DOS version of the MegaRAID Configuration Utility to Linux, but if you use it, don't expect support from LSI Logic or Dell. Since Dell only uses certain LSI Logic controllers in its servers, you might experience varied results when using Dell's MegaRAID management utility. I recommend searching the Web to read about other users' experiences with the controller you purchased to find out if an undocumented bug destroys the information on your disks, or worse.
Download the utility from Matt Domsch's Linux web page (http://www.domsch.com/linux/). Using the utility is very straightforward; it's a statically compiled binary and works the same as the BIOS utility.
Chapter 6. Filesystems
Choosing and properly configuring a filesystem is as important as selecting an appropriate RAID level. ext2 (the Second Extended Filesystem) is the standard Linux filesystem. Many users will be perfectly happy using ext2; it is reliable and can be fine-tuned to meet specific demands of file usage. But while ext2 might be suitable for end users, it doesn't fare as well for large, heavily used filesystems that have extremely large files or thousands of small files. ext2 doesn't provide any way to maintain filesystem integrity through system crashes. Also, ext2 is slowly making way for ext3, a journaling filesystem that I'll cover later in this chapter. Maintaining data integrity and availability has become an essential requirement for all critical systems. After all, this is one of the most recognizable benefits of RAID. Fortunately, there are a wide variety of filesystems for Linux that implement crash recovery and prevention features. These systems are collectively called journaling filesystems, and their main distinction from traditional Unix filesystems is that they don't require filesystem checks after a system crash. I'll cover journaling in greater detail later in this chapter. Even if you aren't interested in journaling filesystems right now, it's a good idea to begin learning about them. It won't be long before your need for increased data reliability or fast recovery may force you to make a change to a journaling filesystem.
There are several alternatives to ext2, in the event that it does not meet your needs. IBM has ported its JFS implementation to Linux. (Like many computer industry terms, JFS has a dual meaning. It refers generically to any journaling filesystem, but also refers specifically to the IBM implementation.) Silicon Graphics has released XFS, the longtime journaling filesystem of the IRIX platform, for Linux. And Hans Reiser has created ReiserFS, a journaling filesystem developed specifically for Linux. Finally, ext3 enhances the features of ext2 by adding journaling capabilities, along with many other features.
This