|
||||
|
|
Linux forums for Web Adminstrators
In This Chapter
Where to Get Commonly Used RPMs
Getting RPMs Using Web-Based FTP
Getting RPMs Using Command-Line Anonymous FTP
Listing Files Associated with RPMs
Which RPMs Will Start Up At Boot Time?
Installing Software Using tar Files
(c) Peter Harrison, www.linuxhomenetworking.com
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
You'll frequently need to install additional software on your Linux server that you didn't think you'd need when you first installed the operating system. This could be because of new business requirements for additional packages or the need to install new administrative tools to make your job easier.
Fedora Linux software is primarily available in RedHat Package Manager (RPM) files. Regular RPM package files are used for installations in which the kernel, or master program, hasn't been customized in any way. This is the usual scenario for most departmental servers. Source RPMs are used when the kernel has been customized to add or drop support selectively for various devices or features for the sake of performance or functionality. The procedure for installing source RPMs involves recompiling source code to fit the needs of these kernel customizations. This makes life easier for the software developer who wrote the package as he or she now has only to create a single package to support all types of customizations. Both package types use standardized commands for installing the software contained inside making RPMs relatively easy to use.
Software developers who want to use a universally recognizable file format across all flavors of Linux also will make their products available as TAR packages. TAR packages are generally more difficult to work with than RPM packages because the archived files within them may or may not need to be compiled and the commands to install the software may vary from package to package. Instructions are usually contained within a file inside the TAR package to help guide the installation.
This chapter focuses on the RPM format, which is the format of choice for Fedora Linux software, but also devotes a small section on TAR packages near the end because they are still very important.
Where to Get Commonly Used RPMs
There are three commonly used sources for RPMs; Fedora distribution CDs; RPMs manually downloaded via a browser, File Transfer Protocol (FTP) client, or the wget utility; and automated downloads using yum. Each of these methods is introduced here, but is covered in greater detail in sections to follow.
RPMs on Your Installation CDs
Installing from your distribution CDs is usually easier than having to download files from a remote Web site, but they are never up to date for very long. We discuss using this method in more detail later.
RPMs Downloaded from Fedora
The two most common ways of getting RPMs from Fedora are by manually using FTP or a Web browser, or by using the automatic yum utility.
Using FTP or Your Web Browswer
You can download RPMs from Fedora using the following link:
http://download.fedora.redhat.com/pub/fedora/linux/core/
You also can use FTP to download Fedora from the download.fedora.redhat.com site. Start your search in the /pub/fedora/linux/core/ directory and move down the directory tree. If you're new to FTP, don't worry, it's explained later.
Remember that Fedora RPMs may not work on RedHat operating systems.
Using yum
The yum program is installed on Fedora systems by default. It enables you to keep the versions of your software up to date by downloading the required packages with the option of installing them afterwards. This is discussed in more detail in a later section.
RPMs Downloaded From rpmfind.net
RedHat and Fedora have only their approved software on their sites. A good general purpose source for additional software is www.rpmfind.net.
This site offers packages for a wide variety of software in various formats for many of the popular versions of Linux. It has an easy-to-use search function that enables you to locate the software you need quickly. Most downloads can be done using HTTP by either clicking on the link or using the wget command, which is explained later.
Note: Always remember to select the RPM that matches your version of Linux. Installing an incorrect version could adversely affect your system.
Getting RPMs Using Web-Based FTP
There are numerous Web sites from which you can download Fedora RPM software; two of the most popular are the Fedora Web site and rpmfind.net. The methods used to get the software from either site are simple, but different enough to be treated separately.
Let's say you are running Fedora Core 2 and need to download an RPM for the DHCP server application; you can select either of these methods in the following sections to get the software.
Using the Fedora Web site
o Use your Web browser to go to the Fedora link given earlier.
o Go to the pub/fedora/linux/core/2/i386/os/Fedora/RPMS directory
o Click on the dhcp-3.0pl2-6.16.i386.rpm link
o Save the file to your hard drive
Using The RPMfind Web site
o Go to the rpmfind link.
o Type in dhcp in the Sarch box.
o Click the search button.
o Scroll down for the RPM that matches your version of Fedora. The right hand column has the links with the actual names of the rpm files.
o Click the link.
o Save the file to Linux box's hard drive.
It is best to download RPMs to a directory named RPM, or something similar, so you can find them later.
Getting RPMs Using Command-Line Anonymous FTP
The Web based method above transparently uses anonymous File Transfer Protocol (FTP). Anonymous FTP allows you to log in and download files from a FTP server using the username anonymous or the shorter username ftp and a password that matches your e-mail address. This way anyone can access the data. Let's illustrate this with an example of using anonymous FTP to download the SSH package from download.fedora.redhat.com:
1. First we issue the FTP command targeting download.fedora.redhat.com at the command line.
[root@bigboy tmp]# ftp download.fedora.redhat.com
Trying 66.187.232.35...
Connected to download.fedora.redhat.com (66.187.232.35).
220 Fedora FTP server ready. All transfers are logged.
Name (download.fedora.redhat.com:root): anonymous
331 Please specify the password.
Password:
230 Login successful. Have fun.
Using binary mode to transfer files.
ftp> pwd
257 "/"
ftp> ls
227 Entering Passive Mode (66,187,232,35,57,155)
150 Here comes the directory listing.
drwxr-xr-x 3 ftp ftp 4096 Oct 29 15:59 pub
226 Directory send OK.
ftp>
2. After we've logged in, we can use the help command to see what options we have at our disposal.
ftp> help
Commands may be abbreviated. Commands are:
|
! $ account append ascii bell binary bye case cd cdup chmod close cr Delete |
debug dir disconnect exit form get glob hash help idle image lcd ls macdef Mdelete |
mdir mget mkdir mls mode modtime mput newer nmap nlist ntrans open prompt passive proxy |
sendport put pwd quit quote recv reget rstatus rhelp rename reset restart rmdir runique Send |
site size status struct system sunique tenex tick trace type user umask verbose ? |
ftp>
The commands you'll most likely use are listed in Table 6-1:
Table 6-1 FTP Commands
|
Command |
Description |
|
binary |
Copy files in binary mode |
|
cd |
Change directory on the FTP server |
|
dir |
List the names of the files in the current remote directory |
|
exit |
Bye bye |
|
get |
Get a file from the FTP server |
|
lcd |
Change the directory on the local machine |
|
ls |
Same as dir |
|
mget |
Same as get, but you can use wildcards like "*" |
|
mput |
Same as put, but you can use wildcards like "*" |
|
passive |
Make the file transfer passive mode |
|
put |
Put a file from the local machine onto the FTP server |
|
pwd |
Give the directory name on the local machine |
3. By using the Web browsing feature on the Web site ahead of time, I know that the Fedora Core 2 RPMs are located in the pub/fedora/linux/core/2/i386/os/Fedora/RPMS/ directory and will use the cd command to change my directory to there. We can use the ls command to get a listing of files in this directory.
ftp> cd pub/fedora/linux/core/2/i386/os/Fedora/RPMS/
250 Directory successfully changed.
ftp> ls open*
227 Entering Passive Mode (66,187,232,35,58,3)
150 Here comes the directory listing.
...
...
-rw-r--r-- ... ... 184281 Oct 28 23:29 openssh-3.6.1p2-34.i386.rpm
...
...
226 Directory send OK.
ftp>
4. Next we get the file we need and place it in the local directory /usr/rpm. The hash command will print "#" hash signs on the screen during the download.
ftp> hash
Hash mark printing on (1024 bytes/hash mark).
ftp> lcd /usr/rpm
Local directory now /usr/rpm
ftp> get openssh-3.6.1p2-34.i386.rpm
local: openssh-3.6.1p2-34.i386.rpm remote: openssh-3.6.1p2-34.i386.rpm
227 Entering Passive Mode (66,187,232,35,58,25)
150 Opening BINARY mode data connection for openssh-3.6.1p2-34.i386.rpm (184281 bytes).
###################################################################################################################################################################################
226 File send OK.
184281 bytes received in 3.41 secs (53 Kbytes/sec)
ftp>
Note: You can also use wildcards to download the RPMs you need using the mget command. You'll be prompted for each of the matching RPM files. In the next example, we just aborted this download by typing n.
ftp> mget openssh-3.6*
mget openssh-3.6.1p2-34.i386.rpm? n
ftp>
5. Finally we use the exit command to leave FTP.
ftp> exit
221 Goodbye.
root@bigboy tmp]#
The wget command can be used to download files quickly when you already know the URL at which the RPM is located. This is especially convenient if you are logged into your Linux box from another machine running a Web browser. You can browse the download site for the RPM you need, right click on the desired link and select copy shortcut (Windows) or Copy Link Location (Linux). After you have done this, you can then select your SSH/telnet/Linux Terminal login window and type in the command wget URL. Here is an example downloading a DHCP update from Fedora.
[root@bigboy tmp]# wget http://linux.stanford.edu/pub/mirrors/fedora/linux/core/2/i386/os/Fedora/RPMS/dhcp-3.0pl2-6.16.i386.rpm
--17:38:36-- ftp://linux.stanford.edu/pub/mirrors/fedora/linux/core/2/i386/os/Fedora/RPMS/dhcp-3.0pl2-6.16.i386.rpm
=> `dhcp-3.0pl2-6.16.i386.rpm.5'
Resolving linux.stanford.edu... done.
Connecting to linux.stanford.edu[171.66.2.18]:21... connected.
Logging in as anonymous ... Logged in!
==> SYST ... done. ==> PWD ... done.
==> TYPE I ... done. ==> CWD /pub/mirrors/fedora/linux/core/2/i386/os/Fedora/RPMS ... done.
==> PASV ... done. ==> RETR dhcp-3.0pl2-6.16.i386.rpm ... done.
Length: 529,890 (unauthoritative)
100%[===============================>] 529,890 889.12K/s ETA 00:00
17:38:36 (889.12 KB/s) - `dhcp-3.0pl2-6.16.i386.rpm.5' saved [529890]
[root@bigboy tmp]#
The yum automatic RPM update program comes as a standard feature of Fedora Core. It has a number of valuable features:
> You can configure the URLs of download sites you want to use. This provides the added advantage of you choosing the most reliable sites in your part of the globe.
> yum makes multiple attempts to download RPMs before failing.
> yum automatically figures out not only the RPMs packages that need updating, but also all the supporting RPMs. It then installs them all.
Note: Updating packages could cause programs written by you to stop functioning especially if they rely on the older version's features or syntax.
Configuring yum
All the configuration parameters for yum are stored in the /etc/yum.conf file. The three basic sections listed in Table 6-2:
Table 6-2 File Format - yum.conf
|
Section |
Description |
|
[main] |
Contains logging and fault tolerance parameters which can usually be left alone. |
|
[base] |
Contains the URL (ftp:// or http://) of a mirror site that contains the Fedora base configuration RPMs. |
|
[updates-released] |
Contains the URL (ftp:// or http://) of a mirror site that contains updated Fedora RPMs. |
The easiest way to determine the exact URLs to use in the baseurl parameters of the [base] and [updates-released] sections of the file is to go to the http://fedora.redhat.com/download/mirrors.html Web site to get a listing of alternative download sites. Browse the sites to find the correct locations of the files.
o The baseurl URL for [base] would be that of the fedora-version/architecture-type/os subdirectory of your version of Fedora. Make sure there is a "headers" sub-directory here, or else it won't work. There will not be RPMs in this subdirectory.
o The baseurl URL for [updates-released] would be that of the updates/fedora-version/ architecture-type sub-directory of your version of Fedora. Make sure there is a headers subdirectory here, or it won't work. There will be RPMs in this subdirectory..
Here is a sample yum.conf file to update Fedora from one of the mirror sites:
[main]
cachedir=/var/cache/yum
debuglevel=2
logfile=/var/log/yum.log
pkgpolicy=newest
distroverpkg=fedora-release
tolerant=1
exactarch=1
[base]
name=Fedora Core $releasever - $basearch - Base
baseurl=http://mirrors.xmission.com/fedora/core/$releasever/$basearch/os/
[updates-released]
name=Fedora Core $releasever - $basearch - Released Updates
baseurl=http://mirrors.xmission.com/fedora/core/updates/$releasever/$basearch/
Note: yum accepts the use of variables in the configuration file. The $releasever variable refers to the current version of Fedora Core running on your server and the $basearch variable maps to the base architecture of your server which is determined automatically.
Note: It is probably best to select yum update sites that use HTTP instead of FTP. There are a number of reasons for this. FTP firewall rules are more difficult to implement than HTTP, outbound HTTP access to the Internet is often already allowed in offices, and web servers are less likely to have connection limits imposed on them, unlike FTP servers, which often have limits on the number of user logins.
Note: You can list multiple URLs in a baseurl statement like this and yum will try them all. If you use multiple baseurl statements in each section then yum may act strangely, frequently only selecting the last one in the list.
baseurl=url://server1/path/to/files/
url://server2/path/to/files/
url://server3/path/to/files/
Creating Your Own yum Server
An obvious advantage of using yum is that you can use it to update a yum server at your office with the same directory structure of the mirror download sites on the Fedora Web site.
A small desktop PC with about five to six gigabytes of free disk space per distribution should be sufficient to start with for a dedicated small business yum server. Large RPMs are about twenty-five megabytes in size, and they are updated infrequently, so your network load should be minimal on average with an update once or twice a week per server. The problem is timing. There is a yum script file in the /etc/cron.daily directory that runs as a cron job every day at 4:00 a.m. Your yum server could get overwhelmed with simultaneous update requests from all your yum clients. If the load gets too high, then you could move this script to another location and schedule it as a cron job for different times for each server. You can also consider throttling the NIC card of the yum server to 10Mbps as another interim means of reducing the problem. Finally, if these measures don't work, you can upgrade the server. For most small businesses/departments this should not be a major concern and you can use MRTG on the server to get trend data for its network load. MRTG monitoring is covered in Chapter 22.
When established, you can then configure all your Fedora servers to use this local yum server for all updates which will significantly reduce your Internet congestion and the associated bandwidth costs.
yum clients can access the yum server using either FTP or HTTP requests. If you need help in setting these up, Chapter 15 discusses Linux FTP servers and Chapter 20 covers the Apache Web server for HTTP requests.
Note: When setting up an HTTP based yum server, you'll need to enable the viewing of directory structures so that it will be easy for someone to use his or her Web browser to navigate down the directories to double check the location of the yum files.
Before You Start
As of Fedora Core version 3, the yum utility checks the downloaded RPMs against checksum files to help protect against file corruption and malicious forgeries. This is set using the gpgcheck variable in the /etc/yum.conf file. When the value is set to 1, and then checks are done, when set to 0, they are disabled.
#
# File: /etc/yum.conf
#
gpgcheck=1
This is a valuable feature to have but you need to load the checksum files in order for yum to work properly. Please refer to the section on "Signature Keys" later in the chapter before proceeding.
Keeping Your System current with Yum
You can make the installed RPM packages on your system up to date with the latest patches using the yum update command. When used without listing any packages afterwards, yum will attempt to update them all. The yum update package-name command updates only a particular RPM package.
It is always advisable to use yum after installing Linux to make sure the latest versions of software are installed for the sake of improved security and functionality. Here is an example of output of what to expect with yum updating your system.
[root@bigboy tmp]# yum update
Gathering header information file(s) from server(s)
Server: Fedora Core 2 - i386 - Base
Server: Fedora Core 2 - i386 - Released Updates
Finding updated packages
Downloading needed headers
Resolving dependencies
Dependencies resolved
I will do the following:
[install: kernel 2.4.22-1.2166.nptl.i686]
[update: samba-client 3.0.2-7.FC1.i386]
[update: binutils 2.14.90.0.6-4.i386]
...
...
...
Is this ok [y/N]: y
Getting samba-client-3.0.2-7.FC1.i386.rpm
samba-client-3.0.2-7.FC1. 100% |=========================| 128 kB 05:01
...
...
...
Running test transaction:
Test transaction complete, Success!
glibc-common 100 % done 1/127
glibc 100 % done 2/127
Stopping sshd:[ OK ]
Starting sshd:[ OK ]
bash 100 % done 3/127
mozilla-nspr 100 % done 4/127
sed 100 % done 5/127
...
...
...
Completing update for pango - 65/127
Completing update for samba-client - 66/127
Completing update for binutils - 67/127
...
...
...
Completing update for XFree86-font-utils - 127/127
Kernel Updated/Installed, checking for bootloader
Grub found - making this kernel the default
Installed: kernel 2.4.22-1.2166.nptl.i686
Updated: pango 1.2.5-4.i386 samba-client 3.0.2-7.FC1.i386 binutils 2.14.90.0.6-4.i386 XFree86-Mesa-libGLU 4.3.0-55.i386 initscripts
[root@bigboy tmp]#
Note: If you don't want to be prompted to install the files use the yum with the -y switch.
Example of a yum Package Installation
Here is a sample installation of an individual package using yum. In this case the RPM installed is the net-snmp-utils package:
[root@bigboy tmp]# yum -y install net-snmp-utils
Repository updates-released already added, not adding again
Repository base already added, not adding again
Setting up Install Process
Setting up Repo: base
repomd.xml 100% |=========================| 1.1 kB 00:00
Setting up Repo: updates-released
repomd.xml 100% |=========================| 951 B 00:00
Reading repository metadata in from local files
base : ############################################ 2622/2622
primary.xml.gz 100% |=========================| 88 kB 00:00
MD Read : ################################################## 229/229
updates-re: ################################################## 229/229
Resolving Dependencies
--> Populating transaction set with selected packages. Please wait.
---> Package net-snmp-utils.i386 0:5.1.2-11 set to be installed
--> Running transaction check
Dependencies Resolved
Transaction Listing:
Install: net-snmp-utils.i386 0:5.1.2-11
Downloading Packages:
net-snmp-utils-5.1.2-11.i 100% |===================| 6.2 MB 00:48
Running Transaction Test
Finished Transaction Test
Transaction Test Succeeded
Running Transaction
Installing: net-snmp-utils 100 % done 1/1
Installed: net-snmp-utils.i386 0:5.1.2-11
Complete!
[root@bigboy tmp]#
Remember The Following Yum Facts
o You can place a list of packages you never want automatically updated in the [main] section. The list must be separated by spaces. Kernel RPMs may be one of the first sets to go on this list, as in this example:
[main]
exclude=kernel
o yum does its updates using TCP port 80 for http:// update URLs and uses passive FTP for ftp:// update URLs in /etc/yum.conf. This will have importance for your firewall rules.
o More details on configuring yum can be obtained by running the man yum.conf command.
o yum runs automatically each day. The cron file is located in /etc/cron.daily/.
o Don't limit yourself to the default yum.conf URLs because they can become overloaded with requests and make yum perform poorly.
There are generally two ways to install RPM files manually. The first method is by using a file previously downloaded to your hard drive, and the other is to install the RPM from some sort of removable media such as a CD-ROM drive.
Using Downloaded Files
Download the RPMs (which usually have a file extension ending with .rpm) into a temporary directory, such as /tmp. The next step is to issue the rpm -Uvh command to install the package.
The -U qualifier is used for updating an RPM to the latest version, the -h qualifier gives a list of hash # characters during the installation and the -v qualifier prints verbose status messages while the command is run. Here is an example of a typical RPM installation command to install the MySQL server package:
[root@bigboy tmp]# rpm -Uvh mysql-server-3.23.58-9.i386.rpm
Preparing... ####################### [100%]
1:mysql-server ####################### [100%]
[root@bigboy tmp]#
Using CD-ROMs
The underlying steps to install RPMs from CDs are similar to those used when installing from your hard disk. The main difference is that you have to access your CD-ROM drive by mounting it first to the mnt/cdrom directory. Your RPMs will then be located in the CD-ROM's Fedora/RPMs subdirectory. The procedure is as follows:
1. Insert the CD-ROM, check the files in the /mnt/cdrom/Fedora/RPMS directory and then install the RPM.
[root@bigboy tmp]# mount
/mnt/cdrom
[root@bigboy tmp]# cd /mnt/cdrom/Fedora/RPMS
[root@bigboy RPMS]# ls filename*
filename.rpm
[root@bigboy RPMS]# rpm -Uvh filename.rpm
Preparing... ####################### [100%]
1: filename ####################### [100%]
[root@bigboy RPMS]#
2. When finished, eject the CD-ROM
[root@bigboy RPMS]# cd /tmp
[root@bigboy tmp]# eject cdrom
[root@bigboy tmp]#
Note: You can use the rpm command's --aid switch to make it search the CD-ROM for any other RPM dependencies and install them automatically.
Sometimes the packages you want to install need to be compiled in order to match your kernel version. This requires you to use source RPM files:
> Download the source RPMs or locate them on your CD collection. They usually have a file extension ending with (.src.rpm)
> Run the following commands as root:
Compiling and installing source RPMs with Fedora can be done simply with the rpmbuild command
[root@bigboy tmp]# rpmbuild --rebuild filename.src.rpm
Here is an example in which we install the tacacs plus package.
[root@bigboy rpm]# rpmbuild --rebuild tac_plus-4.0.3-2.src.rpm
Installing tac_plus-4.0.3-2.src.rpm
Executing(%prep): /bin/sh -e /var/tmp/rpm-tmp.61594
+ umask 022
+ cd /usr/src/redhat/BUILD
+ cd /usr/src/redhat/BUILD
+ rm -rf tac_plus-4.0.3
+ /usr/bin/gzip -dc /usr/src/redhat/SOURCES/tac_plus-4.0.3.tgz
+ tar -xvvf -
...
...
...
+ umask 022
+ cd /usr/src/redhat/BUILD
+ rm -rf tac_plus-4.0.3
+ exit 0
[root@bigboy rpm]#
The compiled RPM file can now be found in one of the architecture subdirectories under /usr/src/redhat/RPMS directory. For example, if you compiled an i386 architecture version of the RPM it will placed in the i386 subdirectory.
You will then have to install the compiled RPMs found in their respective subdirectories as you normally would.
Sometimes the installation of RPM software doesn't go according to plan and you need to take corrective actions. This section shows you how to recover from some of the most common errors you'll encounter.
Failed Dependencies
Sometimes RPM installations will fail giving Failed dependencies errors which really mean that a prerequisite RPM needs to be installed. In the next example we're attempting to install the MySQL database server application, which fails because the mysql MySQL client RPM, on which it depends, needs to be installed beforehand:
[root@bigboy tmp]# rpm -Uvh mysql-server-3.23.58-9.i386.rpm
error: Failed dependencies:
libmysqlclient.so.10 is needed by mysql-server-3.23.58-9
mysql = 3.23.58 is needed by mysql-server-3.23.58-9
[root@bigboy tmp]#
Installing the MySQL client also fails because it requires the perl-DBD-MySQL package.
[root@bigboy tmp]# rpm -Uvh mysql-3.23.58-9.i386.rpm
error: Failed dependencies:
perl-DBD-MySQL is needed by mysql-3.23.58-9
[root@bigboy tmp]# rpm -Uvh perl-DBD-MySQL-2.9003-4.i386.rpm
error: Failed dependencies:
libmysqlclient.so.10 is needed by perl-DBD-MySQL-2.9003-4
[root@bigboy tmp]#
Strangely enough, the installation of the perl-DBD-MySQL package fails because it needs the mysql client package. To get around this problem you can run the rpm command with the --nodeps option to disable dependency checks. In the next example we install the MySQL client ignoring dependencies, followed by successful installation of perl-DBD-MySQL and mysql-server.
[root@bigboy tmp]# rpm -Uvh --nodeps mysql-3.23.58-9.i386.rpm
Preparing... ####################### [100%]
1:mysql ####################### [100%]
[root@bigboy tmp]# rpm -Uvh perl-DBD-MySQL-2.9003-4.i386.rpm
Preparing... ####################### [100%]
1:perl-DBD-MySQL ####################### [100%]
[root@bigboy tmp]# rpm -Uvh mysql-server-3.23.58-9.i386.rpm
Preparing... ####################### [100%]
1:mysql-server ####################### [100%]
[root@bigboy tmp]#
Note: If all the installation RPMs are located in the same directory, the rpm command can automatically install all the prerequisite RPMs using the --aid option. One of the advantages of using the yum facility is that you don't have to worry about this dependency process as much because the dependency RPMs are always downloaded and installed automatically also.
Signature Keys
Fedora digitally signs all its RPM files, so it's best to import their public encryption key beforehand so that the RPM installation program will be able to verify the validity of the RPM file. This can be done using the rpm command as seen in the next example. It is a good idea to import both the RedHat and Fedora keys:
[root@bigboy tmp]# rpm --import /usr/share/rhn/RPM-GPG-KEY
[root@bigboy tmp]# rpm --import /usr/share/rhn/RPM-GPG-KEY-fedora
[root@bigboy tmp]#
If you don't install the keys you'll get a DSA signature warning that alerts you to the fact that the RPM file might be bogus:
[root@bigboy tmp]# rpm -Uvh dhcp-3.0pl2-6.16.i386.rpm
warning: dhcp-3.0pl2-6.16.i386.rpm: V3 DSA signature: NOKEY, key ID 4f2a6fd2
Preparing... #################################### [100%]
1:dhcp #################################### [100%]
[root@bigboy tmp]#
It is always good to install the key files. If they are not there, the RPMs will install with only a warning message. If the RPM's digital signature doesn't match that in the key file, the rpm installation program also alerts you and fails to install the RPM package at all:
[root@bigboy tmp]# rpm -Uvh dhcp-3.0pl2-6.16.i386.rpm
error: dhcp-3.0pl2-6.16.i386.rpm: V3 DSA signature: BAD, key ID 4f2a6fd2
error: dhcp-3.0pl2-6.16.i386.rpm cannot be installed
[root@bigboy tmp]#
Signatures are therefore useful because they help protect you against tampered and otherwise corrupted RPMs being installed.
The rpm -qa command will list all the packages installed on your system
[root@bigboy tmp]# rpm -qa
perl-Storable-1.0.14-15
smpeg-gtv-0.4.4-9
e2fsprogs-1.27-9
libstdc++-3.2-7
audiofile-0.2.3-3
...
...
...
[root@bigboy tmp]#
You can also pipe the output of this command through the grep command if you are interested in only a specific package. In this example we are looking for all packages containing the string ssh in the name, regardless of case (-i means ignore case)
[root@bigboy tmp]# rpm -qa | grep -i ssh
openssh-server-3.4p1-2
openssh-clients-3.4p1-2
openssh-askpass-gnome-3.4p1-2
openssh-3.4p1-2
openssh-askpass-3.4p1-2
[root@bigboy tmp]#
Note: You could use the rpm -q package-name command to find an installed package because it is much faster than using grep and the -qa switch, but you have to have an exact package match. If you are not sure of the package name and its capitalization, the latter method is probably more suitable.
Listing Files Associated with RPMs
Sometimes you'll find yourself installing software that terminates with an error requesting the presence of a particular file. In many cases the installation program doesn't state the RPM package in which the file can be found. It is therefore important to be able to determine the origin of certain files, by listing the contents for RPMs in which you suspect the files might reside.
Listing Files for Already Installed RPMs
This can be useful if you have to duplicate a working server that is already in a production environment. Sometimes the installation of an application fails on the new server due to the lack of a file that resides on the old one. In this case you need to know which RPM on the old server contains the file.
You can use the -ql qualifier to list all the files associated with an installed RPM. In this example we test to make sure that the NTP package is installed using the -q qualifier, and then we use the -ql qualifier to get the file listing.
[root@bigboy tmp]# rpm -q ntp
ntp-4.1.2-0.rc1.2
[root@bigboy tmp]# rpm -ql ntp
/etc/ntp
/etc/ntp.conf
/etc/ntp/drift
/etc/ntp/keys
...
...
...
/usr/share/doc/ntp-4.1.2/rdebug.htm
/usr/share/doc/ntp-4.1.2/refclock.htm
/usr/share/doc/ntp-4.1.2/release.htm
/usr/share/doc/ntp-4.1.2/tickadj.htm
[root@bigboy tmp]#
Listing Files in RPM Files
Sometimes you make a guess and download what you think is the RPM with the missing file. You can use the -qpl qualifier to list all the files in an RPM archive to make sure before installing it:
[root@bigboy updates]# rpm -qpl dhcp-3.0pl1-23.i386.rpm
/etc/rc.d/init.d/dhcpd
/etc/rc.d/init.d/dhcrelay
/etc/sysconfig/dhcpd
/etc/sysconfig/dhcrelay
...
...
...
/usr/share/man/man8/dhcrelay.8.gz
/var/lib/dhcp
/var/lib/dhcp/dhcpd.leases
[root@bigboy updates]#
Listing the RPM to Which a File Belongs
You might need to know the RPM that was used to install a particular file. This is useful when you have a suspicion about the function of a file but are not entirely sure. For example, the MySQL RPM uses the /etc/my.cnf file as its configuration file, not a file named /etc/mysql.conf as you'd normally expect. The following example confirms the origin of the /etc/my.cnf file.
[root@zippy tmp]# rpm -qf /etc/my.cnf
mysql-3.23.58-9
[root@zippy tmp]#
The rpm -e command will erase an installed package. The package name given must match that listed in the rpm -qa command because the version of the package is important.
[root@bigboy tmp]# rpm -e package-name
Which RPMs Will Start Up At Boot Time?
The best way to view and configure which RPMs will start at boot time is by using the chkconfig command with the --list switch. A more detailed explanation will be provided in Chapter 7, which covers the Linux boot process.
Installing Software Using tar Files
Another popular software installation file format is the tar file, which can frequently be obtained from the Web sites of software developers, and online software libraries such as www.sourceforge.net.
The Linux tar command is used to archive files and typically have a .tar file extension in the file name. These files are also frequently compressed in the gzip format, and when they do, their file extensions will end with .tar.gz or .tgz. The commands to extract the data from either type are similar. When a tar file is uncompressed, the command to extract the data is tar -xvf filename.tar. When the archive is compressed, the command to use is tar -xzvf filename.tar.gz.
The tar file installation process usually requires you first to uncompress and extract the contents of the archive in a local subdirectory, which frequently has the same name as the tar file. The subdirectory will usually contain a file called README or INSTALL, which outlines all the customized steps to install the software.
Here are the initial steps to take to install tar-based software:
6. Issue the tar command to extract the files.
[root@bigboy tmp]# tar -xvzf linux-software-1.3.1.tar.gz
linux-software-1.3.1/
linux-software-1.3.1/plugins-scripts/
...
...
...
linux-software-1.3.1/linux-software-plugins.spec
[root@bigboy tmp]#
This creates a subdirectory with the installation files inside.
[root@bigboy tmp]# ls
linux-software-1.3.1 linux-software-1.3.1.tar.gz
[root@bigboy tmp]#
7. Use the cd command to enter the subdirectory and follow the directions listed in the INSTALL and README files:
[root@bigboy tmp]# cd linux-software-1.3.1
[root@bigboy linux-software-1.3.1]# ls
COPYING install-sh missing plugins
depcomp LEGAL mkinstalldirs plugins-scripts
FAQ lib linux-software.spec README
Helper.pm Makefile.am linux-software.spec.in REQUIREMENTS
INSTALL Makefile.in NEWS subst.in
[root@bigboy linux-software-1.3.1]#
Software installation with tar files can be frustrating, frequently requiring the installation of other supporting tar files, each with its own customized installation commands. RPMs, with the single standardized command format, are usually easier to use and may be the better method to use for newer Linux users.
This is just the beginning. If the software you install is intended to make your Linux machine permanently run an application such as a Web server, mail server, or any other type of server you have to know how to get the software activated when the system reboots. This is covered in Chapter 7. Subsequent chapters cover the use, configuration, testing, and troubleshooting of many of the most popular Linux server applications used today.
