Self Test

The following questions will help you measure your understanding of the material presented in this chapter. Read all the choices carefully because there might be more than one correct answer. Choose all correct answers for each question.

1.	What is the minimum number of disks required for a RAID 5 system? Two Five Three One
2.	Consider an e-commerce application with mission-critical data. While configuring the volumes you do not need to be concerned about performance, but you cannot afford to lose data. How would you configure your disk set using the SVM? Stripe Mirror RAID RAID 3
3.	Consider an application in which performance (including read/write performance) is most important. In configuring the volumes you do not need to be concerned about possible data loss. How would you configure your disk set using the SVM? Stripe Mirror RAID 5 RAID 3
4.	Which of the following provides data redundancy by using parity? RAID1 RAID 1 + 0 RAID 5 Hot spare
5.	What is the maximum limit on the number of soft partitions you can create on a disk? Five One Eight No limit
6.	When a slice in RAID 1 or RAID 5 volume fails, it is automatically replaced by a spare slice that stands ready to replace any failed slice. A collection of such spare slices is called: Hot spare pool Spare buffer Redundant slices Disk set
7.	Which of the following is the default read policy for a RAID 1 volume? Parallel First Geometric Round robin
8.	What is the maximum limit on the number of logical volumes that SVM can support on a disk set? Depends on the number of disks in the set 8192 8 No limit
9.	Which of the following volumes write data on components simultaneously? Stripe volume Concatenation volume Logical volume Serial volume
10.	Which of the following commands would you use to add a second submirror (after creating it) to the mirror (RAID 1 volume)? addmirror metainit metattach attachmirror
11.	What is the maximum number of submirrors that SVM can support in a RAID 1 volume? Four Two One No limit

Answers

1.	þ C. RAID 5 implements distributed (striped) parity with striping. You need a minimum of two disks for striping. Because the parity information cannot stay on the same disk where the corresponding data is; therefore, for the parity to be striped you need three disks at minimum. ý A is incorrect because the minimum number of disks you need for RAID 5 is three, not five. B is incorrect because two disks are enough just for striping the data or the parity, but not both. D is incorrect because with just one disk you cannot implement any RAID level.
2.	þ B and C. Both Mirror and RAID 5 volumes support fault tolerance by providing data redundancy. ýA is incorrect because striping does not provide data redundancy, and D is incorrect because RAID 3 is not supported by SVM.
3.	þ A. Striping helps improve performance because it enables multiple controllers to read/write data in parallel. ý B is incorrect because data must be written to more than one component, and C is incorrect because it takes time to calculate parity information. D is incorrect because RAID 3 is not supported by SVM.
4.	þ C. RAID 5 provides data redundancy by calculating and storing the parity information about the data. When the data is lost, it is reconstructed from the parity information. ý A and B are incorrect because RAID 1 and RAID 1+0 provide data redundancy by mirroring and not by parity. D is incorrect because a hot spare is a slice that automatically replaces a failed slice, and it is useful only when data redundancy is already present.
5.	þ D. You can create as many soft partitions as you like; disk space is the only limit. ý A and B are incorrect because there is no such limit as one or five. C is incorrect because the limit of eight partitions applies to a Solaris disk in a standard partition environment, not to soft partitions.
6.	þ A. A slice that stands ready to replace a failed slice is called a hot spare, and a collection of such slices is called a hot spare pool. ý B and C are incorrect because there are no such terms as spare buffer or redundant slices in SVM. D is incorrect because a disk set is a group of disks that holds logical volumes and is not just a collection of hot spares.
7.	þ D. Round robin is the default read policy for a RAID 1 volume. Reads are spread across all submirrors in a round-robin order to balance the read load. ý A is incorrect because parallel is a write policy and not a read policy. B and C are incorrect because First and Geometric are valid ready policies but neither of them is default.
8.	þ B. SVM can support up to 8192 logical volumes per disk set. ý A is incorrect because the number of disks in the set may limit how many volumes you will practically be able to create, but it does not determine the maximum number allowed by SVM. C is incorrect because 8 is the maximum limit on the number of partitions you can create on a standard Solaris disk, not the limit on the number of logical volumes. D is incorrect because there is a limit on the number of logical volumes SVM can support, and that limit is 8192.
9.	þ A. RAID 0 stripe volume writes data on components simultaneously and thereby provides performance improvement. ý B is incorrect because a RAID 0 concatenation volume writes data sequentially. C is incorrect because not all logical volumes write data across components simultaneously. D is incorrect because there is no such volume as serial volume.
10.	þ C. Once a one-way mirror is in place, all the subsequent submirrors should be attached to the mirror by using the metattach command. ý A is incorrect because the metainit command is used to create a submirror, not to attach it. B and D are incorrect because there are no such commands as addmirror and attachmirror.
11.	þ A. SVM supports up to four submirrors in a RAID 1 volume. ý B is incorrect because two is the minimum number (not the maximum number) of submirrors a RAID 1 volume must have. C is incorrect because just one submirror will not make a RAID 1 volume. D is incorrect because there is a limit on the maximum number of submirrors SVM can support, and that limit is four.