Linux Software Management with yum

Linux Software Management with yum

Linux Software Management with yum 150 150 Roderick Derks

In the Linux world, package management systems such as the RPM Package Manager (RPM) and Debian packages have become standard features with most distributions. Package management systems enable easy installs, de-installs, upgrades, and verification of both simple and complex software packages.

The basic tools for both RPM and Debian packages, though, work on individual package files. As Linux distributions have become larger and more complex and as the need for frequent and even automated updates has grown, the limitations of RPM and Debian packages have become apparent.

To fill this void, distribution maintainers have developed tools that either build on existing package systems or scrap them entirely in favor of new systems. Gentoo looked to FreeBSD for the inspiration for its Portage package management, which is network- and source code-based. Debian developers created the Advanced Package Tools (APT) to add network capabilities to Debian packages. In addition to an RPM port of APT, the RPM world has the Yellow Dog Updater, Modified, or Yum, which is similar to APT in basic features and design goals, although it differs in many details.

If you have an RPM-based system, investigating Yum (or APT, if you prefer) is worthwhile. Yum allows you to quickly and easily update the software on your system using network repositories or install entirely new packages.

This article is a copy from, written by Roderick W. Smith. 🙂 Another Roderick..

Like APT and Portage, Yum is a network-based package management system. You give it the network address of a repository that holds RPMs for your distribution along with appropriate Yum metadata. Thereafter, Yum can quickly check for new and upgraded packages, install packages by name without tedious CD-shuffling, and even update your entire system. Updating from one version of your distribution to another is also possible, but can be tricky.

Obtaining and Installing Yum

The first challenge, of course, is to obtain and install Yum. You can download the software from You can also, of course, check the packages that ship with your distribution. (In fact, using a Yum package provided with your distribution is preferable, if one is available. A distribution-provided package is likely to be at least partially configured already and may include distribution-specific tweaks and fixes.)

The Yum site provides links to several different versions the software. As the magazine goes to press, the latest unstable version is 2.9.1, while the latest stable version is 2.6.1. Using the latest stable version is highly recommended. (If you’re using an older distribution, though, you might need to use an earlier version. Yum depends on other software, which might not be recent enough if you’re using an older distribution.)

If you need to use the package provided on the Yum site, be aware that it ships as a tarball (with a .tar.gz filename extension) and as a source RPM (with a .src.rpm filename extension). Binary packages aren’t available. Thus, to install Yum, you must first create a binary RPM (assuming you want to use RPM to manage the Yum package itself). You can do this with rpmbuild:

 # rpmbuild ––rebuild yum-2.6.1-1.src.rpm

Of course, change the version number if you download another version of Yum. The result should be a package in the /usr/src/redhat/RPMS/noarch directory, or something similar. (The redhat directory name will most likely be something else on distributions other than Red Hat or Fedora Core.) Yum is a set of Python scripts, so it’s not architecture-dependent — hence the noarch in the directory path. The final RPM also contains a noarch code in its filename, such as yum-2.6.1-1.noarch.rpm.

Install the package using RPM in the usual way:

 # rpm –Uvh yum-2.6.1-1.noarch.rpm

If all goes well, the Yum utility is now installed. If you see “failed dependencies” or other errors, you should probably track down the missing packages and install them. Package-location sites such as can be very helpful in locating missing packages. You may also run into installation errors if you try to install the latest version of Yum on a less-than-recent distribution. In that case, it may be easier to install an older version of Yum rather than try to track down the missing dependencies.

Configuring Yum

Yum uses the /etc/yum.conf file for configuration. This file contains configuration blocks, each of which begins with a name in brackets, such as [main] or [base]. After the block name come lines that consist of configuration option names, equal signs, and the options themselves, as in:


As with configuration files that use a similar format, you can change options by editing the option values following the equal signs. In most cases, you shouldn’t need to edit the entries in the [main] section of the file, which set global configuration options such as the location of working directories, the debug level, and so on.

One possible exception is the distroverpkg option, which sets the name of a package that contains the distribution version information and additional configuration files (described shortly). This option defaults to redhat-release, but you may need to change this value to suit your distribution. For instance, if you’re running Fedora, it should be set to fedora-release, and for SuSE, it should be suse-release.

In order for the distroverpkg option to be of any use, of course, you must have an appropriate package installed. This package must be for your distribution! If you try to use, say, a SuSE distribution version package on a Mandrake system, you could damage your system so badly that you’ll need to re-install the operating system from scratch. Thus, it’s imperative that you locate an appropriate package. If your distribution ships with Yum, try looking for a package whose name ends (before the version number) in -release. If your distribution didn’t ship with Yum, you’ll have a harder time with this task. Your best bet may be a Web search; try searching on your distribution’s name and “Yum.”

Additional sections of the yum.conf file provide information on Yum repositories that hold packages for the base distribution (that is, the distribution as originally installed) and updates to the distribution (mainly bug fixes). These entries are often shunted into files with .repo filename extensions in the /etc/yum.repos.d directory. This directory often holds multiple files, each devoted to a particular subset of packages, such as the main package set and packages without official support. These files, if they’re present, typically come as part of the package pointed to by the distroverpkg option.

Try checking the /etc/yum.repos.d directory and perusing the files there. The files should contain baseurl and perhaps mirrorlist options, which point to URLs that in turn enable Yum to find files. These URLs often include the variable $releasever in their names, which enables Yum to find your files even when you upgrade your distribution to a new version. You’ll also see enabled lines, with values of 0 to tell Yum not to use a repository or 1 to tell Yum to use it. The default configuration should work well in most cases, but if you want to have Yum check extra repositories, you might need to change the enabled option for some of them.


With Yum installed and configured, you can begin using it. You might want to begin by typing the following command:

 # yum check-update

This command checks your system against the Yum repositories configured in /etc/yum.conf and /etc/yum.repos.d/. The output of this command can be quite verbose. The command’s speed depends on the speed of your network connection and the load at the Yum repository sites. If your system needs no updates, the command returns a value of 0; if any packages need updates, the return value is 100 and the program displays a list of the packages that require updates. This command doesn’t actually download or install any packages, though.

yum check-update is a safe way to begin using yum; it shows you what the program will do when you use it to perform a system upgrade, but without actually modifying any installed packages. If you see error messages or spot upgrades you didn’t anticipate or don’t want, you can take measures to avoid the problems. For instance, you can upgrade individual packages rather than your whole system, as described shortly.

To upgrade all the packages on your system, you can type either of the following commands:

 # yum -y update
# yum -y upgrade

These commands have identical effects, except that the upgrade command activates logic in Yum for handling obsolete packages. (This same feature can be activated by using the ––obsoletes option along with update.) This obsolete package handling is most important when you’re performing a major upgrade, such as upgrading from one version of your distribution to the next one.

These examples use the –y option, which tells Yum to automatically provide a y (“yes”) response to questions asked by Yum. Because “yes” is normally the response you’d give, this can speed up the process of using Yum, but it does increase the risk of problems arising as a result of incorrect automatic “yes” responses. You can change the default by editing the assume-yes option in /etc/yum.conf.

Be aware that system updates and upgrades can download many packages, and some of them can be quite big. Thus, you might want to be careful about performing such updates if you’ve got a slow network connection or if you have little time to spare nursing Yum through the update process. With luck, there will be no problems, but you can never really be sure of that — packages can fail to download or install for any number of reasons, so you might need to re-run the command, remove conflicting packages, or otherwise fix problems.

If you don’t want to update all the packages on your system, you can update just one, as well as any dependencies that need updating, by specifying its name:

 # yum update target-package

This command updates target-package and its dependencies, but nothing else. If you want to see what dependencies might be updated, use check-update rather than update to have Yum report all the packages it will update. You can update multiple packages by specifying them all on one command line.

Installing Software with Yum

As Linux distributions have grown in size, one of the annoyances that’s grown with them is the need to swap CD-ROM’s to locate packages for upgrades. You can avoid this problem by using Yum to grab new packages from Yum repositories on the Internet. To do so, use the install subcommand and name your package:

 # yum install target-package

This command installs target-package on your computer, provided that a package by that name is present in the Yum repositories specified in your Yum configuration file.

In addition to the target package itself, this command installs any packages upon which it depends, if they’re not already installed. Thus, this command can end up installing quite a few packages. You should be aware of this possibility, particularly if your network link is slow.

Yum and x86-64 Systems

If you’re using an x86-64 system rather than a PowerPC, 32-bit x86, or other platform, you should be aware of an extra factor with Yum: By default, Yum install both 32-bit and 64-bit versions of a package, if both are available. In other words, if you type yum install target-package, Yum will look for and install both target-package- version.i386.rpm and target-package- version.x86_64.rpm.

If you know that you only need the 64-bit version of the package, you can force Yum to install only that version by appending the platform specification after a dot, as in yum install target-package.x86_64.

Most distributions make 32-bit versions of packages available for x86-64 versions of their distributions for just a few packages. Thus, you shouldn’t need to specify the architecture very often. Indeed, you should think hard before doing so; chances are if both versions are in the repository, it’s because the package is a library or other package that’s depended upon by other packages that are available only in 32-bit form.

Cleaning Up After Yum

Yum works, in part, by downloading RPMs from repository sites and then installing those packages. By design, though, Yum doesn’t clean up after itself– after installation, the package files remain behind in the Yum directory tree, /var/cache/yum. This enables you to quickly re-install a package if you delete it and then change your mind; however, tit also consumes a great deal of disk space. If you need to get that disk space back, you can do so by telling Yum to do some housecleaning:

 # yum clean packages

This command cleans out the package files, which are generally the biggest disk space hogs. Yum also maintains package header files and other data, though. If Yum begins behaving strangely, you might want to use headers, metadata, cache, or dbcache in place of packages in the preceding command. These options clean the specified types of data, in case they become corrupted or bloated.

Summing Up

All in all, Yum is a great tool for simplifying maintenance of RPM-based systems. By adopting a network-based package management tool, you can more easily keep up with bug fixes (especially security-related bug fixes), install new software, and so on.

This column has described the most important and useful Yum features; however, it supports additional options. You might want to consult the Yum man page, as well as the man page for /etc/yum.conf, to learn more about this useful tool.

Roderick Derks

Liefhebber van fietsen, van het oplossen van IT puzzels, en van het delen van informatie om anderen te helpen.

All stories by:Roderick Derks

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Your Name (required)

    Your Email (required)


    Your Message

      Your Name (required)

      Your Email (required)


      Your Message