Certification Summary

The Redundant Arrays of Inexpensive Disks (RAID) technology takes disk space management from a single disk to multiple disks called a disk set and provides fault tolerance, and, in some cases, performance improvement. RAID 0 provides only performance improvement by enabling the disk controllers to write on stripes simultaneously. RAID 1 provides fault tolerance by duplicating data, and RAID 5 provides fault tolerance by saving the parity information about the data. Additional fault tolerance is provided by a hot spare, which is a slice that stands ready to automatically replace a failed slice.

A logical volume is a collection of physical disk slices that can span across multiple disks. RAID technology can be applied at volume level. The volumes are managed in Solaris by Solaris Volume Manager (SVM), which supports RAID 0, RAID 1, RAID 5, RAID 0 + 1, and RAID 1 + 0 volumes. A RAID 1 volume is also called a mirror, and it is composed of stripe (RAID 0) volumes called submirrors. These submirrors are replicas of each other. You can create a mirror of an existing file system, but you cannot include an existing file system in a RAID 5 volume because the initialization will erase the data. The configuration and status information about the disk sets and about volumes and hot spares in a disk set is contained in a database called a state database. State databases are also managed by SVM.

One of the main benefits provided by the RAID volumes is protection against faults such as a disk crash. However, no data protection method will work if the access to data is not controlled properly. We covered the basic access control in Chapter 7. In the next chapter, we explore advanced methods that provide a more granular approach to access control.

Inside the Exam

Comprehend

• Because a hot spare must be synchronized with the current data, you cannot use the hot spare technology in systems in which redundant data is not available, such as RAID 0.

• In order to provide fault tolerance, each state database replica should reside on a separate disk. For the majority consensus algorithm to work, you must create at least three replicas for the state database when you configure your disk set.

• Because data is written simultaneously on multiple components in a stripe volume, you cannot dynamically change the interlace (data segment size) value. You can dynamically add components to a concatenation volume because data is written on them sequentially.

Look Out

• Striping provides performance improvement but not fault tolerance.

• A soft partition included in a logical volume cannot be directly accessed by an application.

• A. component that already contains a file system (that you don't want to lose) must not be included in a RAID 5 volume creation, because doing so will erase the data during initialization.

• You can include an existing file system in a RAID 1 volume.

• If you mirror a file system that cannot be unmounted, you must reboot your system after creating submirrors and before attaching the submirrors. If you can unmount the file system, then just unmount it and remount it after creating the submirrors.

Memorize

• You can assign a hot spare pool to multiple submirror volumes (RAID 1), or multiple RAID 5 volumes, but a given volume can be associated with only one hot spare pool.

• Stripe volume, concatenation volume, and concatenated volume are all RAID 0 volumes.

• SVM supports up to four submirrors in a RAID 1 volume called mirror.

• While building a mirror (RAID 1 volume), you create the submirrors with the metainit command, and you must attach the second submirror (and any subsequent submirror) to the mirror by using the metattach command, because without the metattach command the submirror will not be resynchronized.

• SVM has the capability of supporting a maximum of 8192 logical volumes per disk set, However, its default configuration is for 128 logical volumes per disk set.

• The default interlace value (size of data segments in stripe volumes and RAID 5 volumes) is 16KB.