Review Questions and Answers

1.	You are the Solaris administrator for your company. You want to create a fault-tolerant solution for your server named filesrv1. Your objectives are to provide data redundancy for all critical data, including the root (/) and /usr file systems, and to keep costs to a minimum. Which of the following is the best solution for your situation? Create a mirrored volume with two submirrors for the root (/) and /usr file systems. Create a RAID 5 volume for application and user data. Create two-way mirrored volumes for all file systems on filesrv1. Create one RAID 5 volume, encompassing all hard disks in filesrv1. Create one RAID 5 volume for the root (/) and /usr file systems. Create a second RAID 5 volume for application and user data. Create a mirrored volume with two submirrors for the root (/) and /usr file systems. Create a RAID 0 volume for application and user data.
2.	On your Solaris server, you have several RAID 0, RAID 1, and RAID 5 volumes. For extra fault tolerance, you have created a hot spare pool named hsp001. You believe that one of your hard disks in a RAID 5 volume failed and want to see whether any of your hot spares are being used. Which of the following commands can you use to check the status of your hot spare pool? (Choose all that apply.) metastat hsp001 metadb hsp001 metahs -s hsp001 metahs -i hsp001
3.	You are configuring Solaris Volume Manager for your Solaris server. Which of the following files keeps track of the locations of state database replicas? /etc/lvm/md.cf /etc/rc2.d/S95svm.sync /kernel/drv/md.conf /etc/lvm/mddb.cf
4.	You are the Solaris administrator for your company. You have just had a hard disk crash. The disk that crashed contained one of your five state database replicas. Solaris Volume Manager needs to know that the replica is no longer available, because you are replacing the device. Which of the following commands can you use to delete the state database replica? metadel metainit metadb metadetach metareplace
5.	You are the Solaris administrator for your network. You are configuring Solaris Volume Manager for first-time use on your Solaris server. When you try to create a RAID 5 volume, Solaris tells you that you must first create state database replicas. Your server has six hard disks. At a minimum, what is the recommended number of state database replicas you should create? One Three Six Seven It depends on the number of slices you have.
6.	You are the Solaris administrator for your company. Your file server SolData1 has a mirrored volume named d10, composed of the submirrors d11 and d12. You bring d12 offline to perform a backup. You know that when you bring d12 back online, it will need to be resynchronized. Which of the following statements accurately describe what will be resynchronized when d12 is brought back online? The entire d12 volume. Only the regions on d12 that fail a parity check test. Only the dirty regions on d12. Only the regions on d12 that are not controlled by the local state database replica.
7.	You are creating a fault-tolerant solution for your Solaris server. You want to ensure that if a single disk or slice fails, your data will still be available online. Which of the following volume types protect your data if one device fails? (Choose all that apply.) RAID 0 RAID 1 RAID 5 Soft partitions Transactional volumes Hot spare pools
8.	You are the Solaris administrator for your network. One of your file servers has three hard disks. You want to create a total of 25 volumes on this file server. How can this be accomplished? You cannot create that many volumes on this computer with its current configuration. Create eight slices on all three hard disks, for a total of 24 volumes, and leave extra space on the last hard disk. Concatenate the last volume on the hard disk into two volumes, giving you a total of 25 volumes. Use disk sets to create 25 virtual slices across the hard disks. Use soft partitions to create 25 virtual slices across the hard disks.
9.	You have recently implemented a RAID 1 volume on one of your file servers. Which of the following are Solaris read policies for RAID 1 volumes that you can choose from? (Choose all that apply.) First Random Geometric Round robin Default
10.	You are the Solaris administrator for your network. You want to create a RAID 0 striped volume containing disk slices c0t2d0s6 and c0t3d0s6. Which of the following commands should you use to accomplish this? metainit d0 1 2 c0t2d0s6 c0t3d0s6 metainit d0 -r 1 2 c0t2d0s6 c0t3d0s6 metainit d0 2 1 c0t2d0s6 c0t3d0s6 metainit d0 -r 2 1 c0t2d0s6 c0t3d0s6
11.	You are the Solaris administrator for your network. You have just increased the size of your RAID 5 volume by adding additional slices. How do you get the file system that resides on that RAID 5 volume to recognize the additional space without losing the existing data? Back up existing data. Delete the existing file system and re-create it. Restore data from backup. Back up existing data. Use the growfs command to expand the existing file system. Back up existing data. Use the growfs command to expand the existing file system. Restore data from backup. Back up existing data. Use the Solaris Volume Manager to expand the existing file system.
12.	You are the Solaris administrator for your company. You have implemented a RAID 5 volume on your server. One of the hard disks in the RAID 5 volume recently crashed. You replaced the hard disk and rebooted Solaris. What do you need to do to resynchronize the RAID 5 volume? From a command line, run the metasync command. From a command line, run the metasynch command. Do nothing; Solaris Volume Manager will resynchronize the volume automatically. RAID 5 volumes cannot be resynchronized. You will need to delete the volume and then re-create it with the new component.
13.	You are the Solaris server administrator for your company. You have created a mirrored volume named d30, which uses a three-way submirror consisting of volumes d31, d32, and d33. You want to make a tape backup from the d33 volume. Which command should you use first so you can make the backup? metadetach d30 d33 metaoffline d30 d33 metabackup d30 d33 metaremove d30 d33
14.	You are the Solaris administrator for your company. In Solaris Volume Manager, you have a RAID volume named d0. You have been instructed to increase the size of d0. Which of the following commands can you use to accomplish this? growfs d0 c0t2d0s2 metadd d0 c0t2d0s2 metaonline d0 c0t2d0s2 metattach d0 c0t2d0s2 metainit d0 c0t2d0s2
15.	You are configuring RAID on your Solaris server. Which of the following statements accurately describes the differences between RAID 0+1 and RAID 1+0? RAID 0+1 is a striped volume that is then mirrored. RAID 1+ 0 is a mirrored volume that is striped. RAID 0+1 is a mirrored volume that is then striped. RAID 1+0 is a striped volume that is mirrored. RAID 0+1 is a striped volume that is expanded to include a concatenated volume. RAID 1+0 is a mirrored volume that expanded to include a concatenated volume. Both RAID 0+1 and RAID 1+0 refer to volumes that are striped and consequently mirrored.
16.	You are creating a RAID 5 volume from three existing slices: c1t0d0s2, c2t0d0s2, and c1t1d0s2. Each of the slices is currently in use and has existing data. You run the command metainit d40 -r c1t0d0s2 c2t0d0s2 c1t1d0s2. Solaris Volume Manager confirms that the d40 volume exists and reports the status as initializing. After the volume changes to okay status, you notice that there is no data on d40. What is the most likely cause of the problem? The volume name you chose, d40, was already in use by another volume. When you created the d40 volume, you should have used the metattach command instead of the metainit command to preserve the existing data. When creating a RAID 5 volume, all existing data on the disks is erased. When you created the d40 volume by using metainit, you should have used volume numbers representing the slices instead of the slice designations. This would have preserved the existing data.
17.	You are the Solaris administrator for your network. You have recently expanded the size of one of your file systems. Now, however, you need to make the file system smaller, so you have room to create another file system on that hard disk. Which of the following should you do to accomplish this? Back up existing data. Use the growfs command to shrink the file system. Create the new file system as needed. Back up existing data. Use the growfs -r command to recursively claim the expanded file system's space. Create the new file system as needed. Back up existing data. Use the shrinkfs command to shrink the file system. Create the new file system as needed. Back up existing data. Delete the existing file system and re-create two smaller ones. Restore data from backup. Create the new file system as needed.
18.	You are the Solaris administrator for your company. One of your file servers has been serving more users recently, and users are complaining that file access times on the server are decreasing. Which of the following can you implement to provide the greatest speed increase to users accessing this file server? RAID 0 striped volume RAID 0 concatenated volume RAID 1 RAID 5 Soft partitions Transactional volumes
19.	You have recently implemented a RAID 1 volume on one of your file servers. Which of the following are Solaris write policies for RAID 1 volumes that you can choose from? (Choose all that apply.) Parity checking Write through Write back Parallel Serial
20.	One of your Solaris servers contains a large number of hard disks. Because of RAID volumes and soft partitions, you need to increase the default number of volumes you can create on that server. Which file do you edit to perform this task? /kernel/drv/md.conf /etc/lvm/md.cf /etc/rc2.d/S95svm.sync /etc/lvm/mddb.cf

Answers

1.	A. The only two fault-tolerant volume types available (by default) in Solaris 9 are RAID 1 (mirrored volumes) and RAID 5 (disk striping with parity). RAID 5 volumes are more cost-effective than mirrored volumes. However, you cannot place the root (/) and /usr file systems on a RAID 5 volume. If you want to protect data on those two volumes, you must use a mirror.
2.	A, D. The two commands you can use are metastat and metahs -i. The metadb command is used to manage state database replicas. The metahs -s switch is used to specify disk sets when creating hot spare pools.
3.	D. The /etc/lvm/mddb.cf file, also known as the metadevice database, keeps track of the locations of state database replicas.
4.	C. The metadb command is used to create, manage, and delete state database replicas. metadel does not exist. metainit creates volumes, metadetach detaches components from a RAID 1 or transactional volume, and metareplace is used to replace devices in a RAID 1 or RAID 5 volume.
5.	B. The recommended minimum number of state database replicas is three. Because your computer has more than three hard disks, it wouldn't be a bad idea to create more (and an odd number is better than an even number). However, three is still the minimum.
6.	C. When a submirror is brought offline, the other submirror still reads and writes data. The new data written is tracked by Solaris Volume Manager, and regions that are not updated (on the offline volume) are called dirty regions. When the d12 submirror is brought back online, the dirty regions need to be synchronized.
7.	B, C. The only two types of volumes that provide fault tolerance (data redundancy) are RAID 1 and RAID 5. RAID 0, soft partitions, and transactional volumes do not provide fault tolerance. Hot spare pools can be used by RAID 1 and RAID 5 volumes for added redundancy; however, hot spare pools in and of themselves provide no fault tolerance, nor are they a volume type.
8.	D. The only way to increase the number of volumes on a hard disk beyond the limit of eight is to use soft partitions. Even with soft partitions, you will never get more than eight slices on one hard disk. However, you will be able to get more than eight logical volumes, which, for all practical purposes, act like slices when it comes to file and application access.
9.	A, C, D. There are three RAID 1 (mirrored volume) read policies in Solaris 9. They are round robin (the default), which reads disks in order, geometric, which assigns logical disk address ranges to components, and first, which always reads the first disk if it's available.
10.	A. The correct syntax for creating a striped RAID 0 volume is # metainit volume_name #_ stripes components_per_stripe component_names -i interlace. The -r switch is used to create RAID 5 volumes.
11.	B. After a volume is expanded, the growfs command can be used to expand the underlying file system. The growfs command does not destroy data, so there is no need to restore data from backup after using growfs.
12.	C. Upon replacement of a failed component, Solaris Volume Manager will automatically resynchronize a RAID 5 volume. It is possible to manually synchronize volumes with the metasync command (metasynch does not exist), but it's not necessary to run the command in this case.
13.	B. The metaoffline command is used to bring submirrors offline. After they are offline, you can use the d33 volume to make a tape backup. After you bring d33 back online (with the metaonline command), it will resynchronize with the rest of the d30 volume immediately.
14.	D. The metattach command is used to expand existing RAID volumes. The growfs command grows file systems, not volumes (although after running metattach, it might be a good idea to expand the file system with growfs). The metadd command does not exist; metaonline brings submirrors online in a RAID 1 volume, and metainit creates new volumes.
15.	A. A RAID 0+1 volume starts off as a RAID 0 volume (typically a striped volume) and then is mirrored. RAID 1 volumes are mirrors, which can then be striped to enhance performance. These types of volumes are called RAID 1+0.
16.	C. When you create a RAID 5 volume, all data that is on the slices used to create the volume will be destroyed. This is because when the RAID 5 volume is created, Solaris Volume Manager needs to calculate parity bits. These parity bits will be used to provide an online backup of data. RAID 5 volumes are created with the metainit command, and you must specify slice names, not volume names, as components.
17.	D. After a file system is grown, it cannot be shrunk. To free up space for another file system, you will have to delete the existing file system and re-create a smaller one.
18.	A. Of the RAID volumes, RAID 0 striped volumes provide the fastest read and write access times. Soft partitions and transactional volumes do not provide a disk speed increase.
19.	D, E. There are two RAID 1 write policies: parallel (the default) and serial. In parallel, all disks are written to at the same time. In serial, the first write must complete before the second write begins.
20.	A. The /kernel/drv/md.conf file is what you're looking for. Inside this file, the field you need to edit is called nmd. Increase the number from 128 to whatever you need it to be. Remember, though, that increasing this value unnecessarily high can slow your system, and the maximum value is 8192.