anaconda has the capability to incorporate updates at runtime to fix any bugs or issues with the installer. These updates are generally distributed as a disk image file (referred to as
updates.img from here on out). The
updates.img can be used in a few different ways.
There are a number of sources for the updates.
Updates from the Network
The easiest and most popular way to use an
updates.img is via the network. This is how almost all updates images you'll see in bug reports and mailing lists are distributed. This does not require you to modify your installation tree at all.
To use this method, simply boot with:
If you have multiple network interfaces, anaconda will first prompt you to select one (unless you have used the
ksdevice= boot parameter). It will then attempt to configure this link using DHCP. If you require other networking configuration, you will need to use various options.
ksdevice= can be used to specify a different network device, and the
ip= option (along with others for gateway, nameserver, and so forth) can be used for static configuration. All anaconda config options are described elsewhere .
If you are making your own
updates.img, just upload it to a web server you have access to and pass the location as above.
Updates from a disk image
You can also put an
updates.img on a block device (either a floppy or a USB key). This can be done only with an ext2 filesystem type of updates.img. For a floppy drive, insert your floppy and then run
dd if=updates.img of=/dev/fd0 bs=72k count=20
to put the contents of the image on your floppy. Then, boot the installer with
and you will be prompted to provide the location of your update disk.
You can also use a USB key or flash media -- just replace
/dev/fd0 with the device that your USB key is at.
Updates from the Tree
If you're doing a CD, hard drive, HTTP, or FTP install you can also put the
updates.img in your tree to be picked up by all installs automatically. Put the file in the
images/ directory. It must have exactly the name
updates.img, even if you received it with a different name.
For NFS installs, there are two options. You can either put the image in
images/ as above or explode the image into the
RHupdates/ directory in your installation tree.
updates.img is only retrieved from the location where stage2 image is pulled from. If you use inst.repo boot option to specify your installation tree, but you also use inst.stage2 boot option with a different location, only the
inst.stage2 location is going to be searched for the
updates.img file, and not the
How to Create an Anaconda Updates Image
If you are working on anaconda or looking at a bug and want to test your own bug fixes, it's easy to create your own
updates.img file. anaconda supports two formats: an ext2 filesystem image and the more common gzip-compressed cpio archive. The automatic tools shipped with anaconda deal in the second form, so that's what will be discussed here.
The easiest way to create an image is to run
$ ./configure $ make updates
from the anaconda source tree. This will package up all the changes to the tree since the last release and create a file named "updates.img" in the top of the tree. Remember to use the correct git branch for the Fedora release you are working on or testing. If you need finer control over this process (like creating an image from an even older release), or you don't want to run ./configure first (the make command will fail unless ./configure has been run), run
by hand. The help screen documents the several options that can be used.
updates.img can include more than just files from anaconda, though. It can also include shared libraries, graphics, other python modules, and certain data files used by anaconda. To add files to an existing image (or create an entirely new one), just do the following:
$ scripts/upd-updates updates.img file1 file2 ...
Note that the placement of files in an image is a little picky. For instance, python modules must be in their proper subdirectory mirroring the layout of
Another way to create an image containing files outside of anaconda is to create the required filesystem structure and compress it manually. For example, let's say you want to overwrite some configuration file in
$ mkdir -p updates/etc/ $ cp my.cfg updates/etc/ $ cd updates $ find . | cpio -o -c | gzip > ../updates.img $ cd ..
How to Examine an Anaconda Updates Image
updates.img files provided by the Fedora project and generated by the makeupdates script are compressed cpio archives. To examine one of these files, use
$ lsinitrd updates.img
To explode one, do the following:
$ mkdir dest $ cd dest $ gunzip -dc /path/to/updates.img | cpio -id
usage: makeupdates [-h] [-k] [-c] [-t TAG] [-o OFFSET] [-p] [-a PATH_TO_RPM [PATH_TO_RPM ...]] [-f ARCH] [-b BUILDDIR] Make Anaconda updates image optional arguments: -h, --help show this help message and exit -k, --keep do not delete updates subdirectory -c, --compile compile code if there are isys changes -t TAG, --tag TAG make updates image from TAG to HEAD -o OFFSET, --offset OFFSET make image from (latest_tag - OFFSET) to HEAD -p, --po update translations -a PATH_TO_RPM [PATH_TO_RPM ...], --add PATH_TO_RPM [PATH_TO_RPM ...] add contents of RPMs to the updates image -f ARCH, --fetch ARCH autofetch new dependencies from Koji for ARCH -b BUILDDIR, --builddir BUILDDIR build directory for shared objects
Including Updates for an Older Installation Image
If your installation image has an older Anaconda (for example you have a Beta image but you want to test all the changes that happened in Anaconda since the image was created), you can use the
-t makeupdates option, together with the Anaconda release tag corresponding to the Anaconda version on your image. Makupdates will then include all changes that were added since the given Anaconda version was released.
How to Find Anaconda Version for an Installation Image
There are multiple ways how to do that:
- switch to TTY1 and check the first line on the screen
- check the first line of the anaconda.log file in /tmp/ during installation
- check the first line of the anaconda.log file in /var/log/anaconda on a system installed with your installation image during installation
- check the version of the Anaconda package in the repository that has been used to generate your installation image
- boot a Fedora installation image
- find what version of Anaconda is installed on the image
- lets say that the image contains Anaconda 22.16-1
- this version corresponds to the anaconda-22.16-1 Git tag
- you can run
git tagin the Anaconda git repository to list all valid tags
makeupdates -twith the tag:
makeupdates -t anaconda-22.16-1
- an updates image containing all changes since the commit tagged
anaconda-22.16-1will be created
Including Changes in C Code
While Anaconda is mostly written in Python, there are a few pieces of C code, mostly in the form of custom GTK Widgets and the isys helper module. The makeupdates ignores changes in C code by default, but by passing the
-c option you can tell it to look for C code canges, recompile the affected modules and include the resulting binaries in the updates image.
Just take not that for the compilation to finish successfully, the host system needs to match the given Installation Image. This is especially important when rebuilding the custom GTK widgets.
So it is for example not possible to use the
-c option on a Fedora 21 system to build an updates image with C code changes for a RHEL7 Installation Image or the other way around.