Using Stork in place of Yum
or Apt
Although Stork includes many advanced features such
as security and automated package installation,
Stork also makes an excellent replacement for the
command-line package installation tools such as yum
and apt.
Stork uses a repository located at
http://stork-repository.cs.arizona.edu and this
repository is already populated with Fedora Core 2
and Fedora Core 4 packages for PlanetLab V3 and V4
nodes respectively. Although this tutorial
concentrates primarily on using these pre-provided
packages, there are mechanisms for adding your own
packages to the repository and we direct you to the
Advanced Usage section
for more information.
Installing Stork on your Slice
[You may skip this section if stork is already
installed and running on your Slice]
In order to use Stork, it must first be installed
on your Slice. We are currently working with the
PlanetLab developers on an automated mechanism for
enabling Stork, but for the time being you will need
to manually run our initscript
to install the stork packages on your nodes. You will
need to do this on each node of your slice that you
want to run Stork.
Running the initscript requires root access, so
be sure to do a 'su root ' if you have not done so
already.
# wget
http://www.cs.arizona.edu/stork/downloads/initscript
# ./initscript
The stork initscript should download the stork
packages to your node and install them. It will also
start up a service called PacMan that automatically
keeps the current version of stork up-to-date for
you.
Once the initscript has completed, you may start
using Stork to install Fedora 2 / 4 packages from
our repository immediately. You do not need to edit
any of the Stork configuration files, trusted
packages files, or keys unless you wish to configure
Stork to install your own custom packages.
Manually managing packages with Stork
The stork client tool can be used to manually
install, remove, and upgrade packages. It does this
through a command-line interface. You must be logged
in as root to use the stork client tool.
Installing Packages:
To install a package on a node,
use the command:
# stork lynx
Removing Packages:
To remove a package from a node, use the option:
# stork --remove lynx
Upgrading Packages:
To upgrade a package from a node, use the option:
# stork --upgrade lynx
Installing specific version numbers:
To install a specific version number of a
package:
# stork --upgrade lynx=1.0.3-11
You may also use >, <, >=, <=
relational operators:
#
stork --upgrade
lynx <=1.0-3-11
The stork client command-line tool has many
options. For a complete reference, see the
Stork Client reference
page.
Notes / Frequently Asked Questions
This section contains information on some
important differences between Stork and Yum, and
some issues that you might experience.
How does Stork choose which package to install?
Generally there may be more than one version of
a package in our repository, as our repository
contains both Fedora 2 and Fedora 4 packages.
Stork has an automatic rule to decide whether your
node is a PlanetLab V3 or V4 node and
automatically prefer the appropriate version of
the Fedora packages.
If you want to read more about how Stork
decides which version of a package to install,
then consult the Advanced
Topics page and read more about Trusted
Packages Files.
Error: "Another copy of stork is already
running..."
Stork automatically starts a service called
PacMan, which is responsible for automatically
installing and updating packages. By default, it
only maintains the Stork packages themselves.
Periodically PacMan will wake up and run. When
PacMan is running, the stork command line tool
will be unavailable. If you get this error
message, then we suggest you wait a few minutes
and try again. You can deactivate PacMan by
executing "/etc/init.d/pacman.init stop". However,
we encourage you not to do so as this will prevent
your Slice from receiving important Stork updates
and bug fixes.
-
Home Top
|