Solaris 9 Sun Certified System Administrator Study Guide

Know the features of a RAID 0 striped volume. RAID 0 volumes are created on two or more hard disks, and are used to increase disk read and write speed. RAID 0 volumes do not provide any fault tolerance.

Know the features of a RAID 1 mirrored volume. Mirrored volumes provide data redundancy. If a mirrored device fails, the "copy" of the data is still available online for use. Implementing RAID 1 requires an investment in hard disk space.

Know the three RAID 1 read policies. The three read policies are round robin, geometric, and first. Round robin, the default, means that each submirror is read in a serial fashion. On the first read, the first submirror is read. On the second read, the second submirror is read, and so on. The geometric read policy attempts to optimize disk access time by assigning each submirror a set of logical disk addresses to read from. The first policy means that the first submirror is always consulted first.

Know the two RAID 1 write policies. The two write policies are parallel and serial. Parallel is the default write policy, and it means that all submirrors are written to at the same time. With the serial policy, the first submirror's write must be completed before the second submirror's write is started.

Know the features of a RAID 5 volume. RAID 5 volumes write the data in stripes, just as a RAID 0 striped volume does. However, RAID 5 also provides fault tolerance by calculating and writing parity information on the hard disks. Although you lose storage capacity (for the parity information), RAID 5 volumes are more ergonomically correct than RAID 1 volumes.

Understand what a volume is. A volume (also known as a metadevice) is a logical unit of slices, acting as one, and managed by the Solaris Volume Manager.

Know what the metadevice database is. The /etc/lvm/mddb.cf file, which contains the list of state database replicas, is the metadevice database.

Know how to grow a file system. Volumes are increased by adding slices. After the volume size is increased, you can grow your file system (to use all of the new space) with the growfs command.

Know what state database replicas are. State database replicas contain the configuration information for volumes created on your computer. For optimal results, you should have at least three state database replicas, located on different hard disks and controllers if possible.

Know how to create volumes from the command line. Volumes are created by using the metainit command.

Know what resynchronization is. When a RAID 1 or RAID 5 component fails and is replaced by a new component, the new component needs to have current data copied to it. This process is called resynchronization.

Know the advantages of using soft partitions. Soft partitions enable you to create more than eight logical volumes per disk. Also, soft partitions can have their size increased (growing the partition) dynamically, without requiring a reboot of the system.

Know what transactional volumes are used for. Transactional volumes are used to log UFS reads and writes. They are being phased out in favor of UFS logging.

Know what hot spare pools are for. A hot spare pool is a collection of unused slices that are standing by in case of a RAID 1 or RAID 5 component failure. When a RAID 1 or RAID 5 component fails, a hot spare is automatically pulled from the pool and used to replace the failed component.

Know what a disk set is. A disk set is a logical collection of volumes and hot spares. Shared disk sets can be controlled by more than one computer, but never more than one computer at any given time.