Yum - Some Simplifying Tricks

If you are using a distro that is able to use YUM and you have a couple of systems, you can save lots of time by:-

This section describes how you can setup one machine and let others use the updates for themselves. It does not describe how to setup your own yum repository(that comes in another document)

So, I have 2 systems. The first called Trophy650 is my general server. It is running Fedora Core 3 and is installed with "everything" (all packages). A cron script runs daily at 01:00 and does a "yum -y update". This not only updates the system but provides the package repository for the other system to use.

The second system is called "TigerCub" and runs a mixture of FC3(X86) and FC3(X86-64). This is an AMD 3200 system. For the X86 distro, it is a subset of the packages used on Trophy650 that is tailored towards running commercial software like Websphere MQ, DB2 & Oracle.

There are two ways I can update TigerCub from Trophy650. Firstly, I could setup Trophy650 as a yum server and point yum on TigerCub at it. This option is going to be covered in another document. The second way is a quick and dirty one but it is also easier to do initially.

When you do a "yum update", the packages etc are downloaded into the tree "/var/cache/yum/updates_released". This has two further subdirectories. 1) headers, 2) packages. These are the key directories that are needed on your target system.

The steps are as follows:-

As you can see, this is a quick and dirty method of ysing yum in this way. It does solve the problem of the so called "rpm dependancy hell".

A far neater way is to setup your server as its own "yum repository" and point the target system at it. This will be covered in another wiki entry along with removing that annoying Red/Blue up2date icon on your FC3 screen.

ManagingPackages/Yum (last edited 2005-03-30 14:32:45 by 152)