Lab Questions

---

Lab 1

1. 

You just got hold of ten new PCs for the Human Resources department from a name brand PC manufacturer and you want to install Red Hat Enterprise Linux on each computer. You want to install Linux on all of them with an optimized set of packages, and you want to do it quickly.

Each of these computers has a standard 3Com network card that you know Linux has support for because you ordered the computers that way. They also each have one big 10GB disk that already contains Windows 98. You do not have time to install Red Hat Enterprise Linux 3 Workstation on each computer manually. What should you do?

Assume you need to configure each computer with static IP addresses.

Answers

1. 

The solution is simple. The details are long. Each of the following steps are high level; for the nitty- gritty details of what you need to do, refer back to discussions in this chapter (as well as Chapter 2). If you're using these instructions to restore a RHEL 3 installation from Part 1, minimize the changes that you need to make to this file.

1. Set up the RHEL 3 Workstation installation files on a network source.

2. Install RHEL 3 Workstation over the network on one computer.

3. Edit the /root/anaconda-ks.cfg file as needed.

4. Windows 98 partitions. (This assumes you've backed up any needed data files first.)

5. Activate the command lines related to creating partitions. They are commented out by default in the anaconda-ks.cfg file.

6. In the Kickstart configuration file, make sure that the line after the install command points through the network to the source for the RHEL 3 Workstation installation files. If you're connecting through NFS, this starts with the nfs command; if you're connecting through FTP or HTTP, this starts with the url command.

7. Use this Kickstart file as a configuration file template for installing Linux on each of your computers.

8. Create a different file for each computer; you'll need to set a different IP address on the same subnet for each file.

9. Set up a different installation floppy for each computer. Each installation floppy will include the customized Kickstart file. You'll probably also need one or more driver floppies. Alternatively, you can set this up based on the boot.iso file. All of these disks can be created from images on your first RHEL 3 installation CD. This process is described in Chapter 2.

10. You can then set up a network installation on each computer. You'll probably also need the RHEL 3 network driver floppy described in Chapter 2. You can use the same driver floppy on each computer.

11. Now insert the appropriate installation floppy in each of your workstations. When prompted, add the driver floppy to load the appropriate network drivers.

12. Your installation should now proceed automatically on each workstation.

13. Save any installation boot and driver disks or CDs that you created in this lab for Part 2.

Lab 2

2. 

This lab requires that you have one of the network installation servers described in Chapter 2 and a test computer where you're willing to reinstall RHEL 3. This exercise erases any operating system that you currently have on a computer! Do not run this lab unless you can afford to lose all files and are willing to reinstall RHEL 3 on this computer. You're going to move one of the critical files out of the installation source directory.

You could perform this lab on a new VMWare machine. Alternatively, you could run this on a computer with a standard configuration, where you can set up an automated reinstallation with a Kickstart configuration file. You've probably just seen this process in the answer to Lab 1.

1. Log into the remote network server with your RHEL 3 installation files.

2. Look at your comps.xml configuration file. It's stored in the RedHat/base/ subdirectory of the installation source. Open it in a text editor or a text reader such as the less command.

3. Browse through the list of packages in XML format. Pay attention to the packages in the Base and Core package groups.

4. Pick a RPM package from one of these groups. If you can't think of one, I've selected the filesystem-2.2.1-3.i386.rpm file.

5. Move that file to the home directory to some other directory on that network installation server. Make sure to use the mv command; the cp command leaves a copy on the installation source.

If you have valuable data on this computer, STOP HERE! The following steps will delete the information at least on the Linux partitions on your computer-and possibly on the entire computer.

1. Return to your test computer, and proceed through the RHEL 3 installation process.

2. At some point, the RHEL 3 installation process will miss the RPM that you moved and stop with a message to that effect.

3. Press CTRL-ALT-F2. You should see a bash prompt. Look in the /mnt/sysimage directory. What is the directory tree in this directory?

4. Navigate to the /mnt/sysimage/root directory. Look at the files in this directory. Which file tells you about the RPMs that have been installed?

5. Run the chroot /mnt/sysimage command. If it doesn't work, your installation stopped before RHEL 3 had a chance to install the bash shell. With the RPM cited, this is the case.

6. Press CTRL-ALT-F3. Take a look at the messages. You should see the partitions you specified, mounted in some way. Where are they mounted?

7. Press CTRL-ALT-F4. Take a look at the messages. You should see various kernel messages. What types of messages do you see?

8. Press CTRL-ALT-F5. Take a look at the messages. You should see various partition messages. What commands do you see in this screen?

9. On the network server, restore the RPM that you moved.

10. Depending on whether you're installing in text or graphical mode, return to your installation screen with the CTRL-ALT-F1 or CTRL-ALT-F7 command.

11. Proceed with the installation of RHEL 3.

Answers

2. 

This lab should be fairly self-explanatory. It provides a specific problem that stops the installation of RHEL 3. It gives you an opportunity to examine the other consoles available during the installation process. Remember, you should not run this lab on any computer where you have valuable data. I've added some commentary which may clarify the steps you need to take in this lab.

The comps.xml file is located in the installation source. If you're using the same NFS installation source that I described in Chapter 2, it's on the /mnt/inst directory. Thus, the comps.xml file is located in the /mnt/inst/RedHat/base directory. If you already have an installed copy of RHEL 3 on an Intel-compatible 32-bit CPU computer, you can find another copy of this file in the /usr/share/comps/i386 directory.

The format of the comps.xml file might be confusing to anyone not familiar with the XML language. However, it's not as complex as it seems. For example, you can find the filesystem RPM in this file in the following format:

<packagereq type="mandatory">filesystem</packagereq>

When you browse higher on this file, you'll see that the filesystem RPM is part of the Core package group.

Once the installation stops at the problem you created, you can browse through the different consoles. The second console includes the bash prompt, even though bash commands aren't installed yet. You should also recognize the standard root directory (/) tree in the /mnt/sysimage directory.

In the third virtual console, you'll see messages that partitions such as /tmp/hda1 or /tmp/sda1 are mounted. For example, a partition mounted on the root directory is mounted on the /mnt/sysimage directory during the installation process. The partitions that you've created through Disk Druid are mounted on a temporary basis until installation is complete.

The kernel messages that you see in the fourth virtual console show hardware and filesystem messages. The fifth virtual console lists the messages you see when the filesystem is formatted.

Lab 3

3. 

What would you do differently from Lab 2, assuming that you have configured a DHCP server for your network?

Answers

3. 

The basic steps are the same. However, if the ten computers are connected to a network with a DHCP server, your task is easier. Modify the aforementioned Kickstart template file. Assuming it's the first Ethernet adapter on each computer, the key command will be network --device eth0 --bootproto dhcp.

Since you're getting network information from a DHCP server, you can use the same installation floppy for each of the workstations. You'll still need the driver floppy. Now insert the appropriate installation floppy in each of your workstations. When prompted, add the driver floppy to load the appropriate network drivers. You should be on your way!

Alternatively, you can create a customized boot CD based on the boot.iso file on the first RHEL 3 installation CD. Add the ks.cfg file when you burn your CD from boot.iso.

Lab 4

4. 

You want to set up a server for these new PCs. You want to configure that computer with the Red Hat Enterprise Linux 3 Server operating system. That server includes five separate SCSI physical hard drives of 10GB each. You'll install RHEL 3 on the first hard drive. How would you configure a RAID 5 array on the remaining hard drives? Assume you'll need 5GB from each of the 10GB hard drives.

Answers

4. 

Now you're configuring a RAID 5 array for your workstations on a RHEL 3 server. With the given configuration, RHEL 3 will be installed on the first hard disk. The remaining four hard disks are available for other purposes. You'll use 5GB of each of those SCSI hard disks to create RAID 5 arrays. It's easiest to do so during the installation process. However, you can still do so after RHEL 3 is installed.

As they are SCSI hard disks, Linux represents them as /dev/sdb, /dev/sdc, /dev/sdd/, and /dev/sde. You can divide the 5GB on each of these hard disks into near equal parts (RAID does not require identical components in its arrays.). If you created 5GB partitions on each hard disk, you could set up a RAID 5 array with the first 3 partitions: /dev/sdb1, /dev/sdc1, and /dev/sdd1. You could then set up the partition on the final drive, /dev/sde1, as a spare. (While you could set up all four partitions as a RAID 5 array, it's useful to have a spare. If one drive fails, the RAID 5 software automatically begins rebuilding data on the spare drive.)

You can create a RAID array after RHEL 3 is installed, using the following basic steps. For detailed steps, refer back to the chapter.

1. Use the fdisk utility to create and size partitions from available space on the SCSI hard disks.

2. With the t switch, set these partitions as usable by RAID by setting them to the Linux RAID autodetect filesystem.

3. Format each partition, set up the array in /etc/raidtab, and use the mkraid -R command to create the RAID device.

4. Format the RAID device to the Linux ext3 filesystem.

5. Save any data from the directory that you want to mount on the RAID array.

6. Mount the directory of your choice to that filesystem, and finally, implement the change in /etc/fstab.

Lab 5

5. 

On that server, you want to configure LVM in the remaining space. How would you set up the Physical Volumes (PVs)? How do you create a Volume Group (VG) from the PVs? Once you've created a VG, how would you set up a Logical Volume? Where would you mount a directory such as /home/flexuser?

Answers

5. 

You can create LVM groups after RHEL 3 is installed, using the following basic steps. For detailed steps, refer back to the chapter.

1. Use fdisk to create partitions from the remaining available space on each of the last four SCSI drives. Make sure to set each partition to the Linux LVM filesystem.

2. Initialize each of these partitions with the pvcreate command.

3. Set up all of these partitions together as a Volume Group (VG) with the vgcreate command.

4. Now set up a Logical Volume from that VG with the lvcreate command. You can set it to a fixed number of Physical Extents or a fixed amount of space.

5. Take the directory of your choice. Back up any data currently on that directory.

6. Mount the directory on the LV. Copy the data from backup.

7. Finally, set up the mount in the /etc/fstab configuration file.

---