Connecting the dots with Solaris and middleware In the middle

This entry addresses publishing and using a reduced installation profile metapackage and the Automated Installer (AI) as a means of simplifying the installation of heavily reduced forms of OpenSolaris.
[Read More]

Since build 126 of OpenSolaris was recently released, it was time to update my Automated Install (AI) server and to verify that the reduced installation profile I had tested using build 125 was still working properly.  Overall, the process was straightforward and trouble free.

Updating the reduced OpenSolaris installation took very little time because only 78 MB was needed to update the 200 or so packages installed on my AI server.  This is a key benefit of using a reduced installation of an OS: you can dramatically reduce update overhead by downloading and applying updates for only those packages that matter to your particular application.  It's also a testament to the pkg(5) system in that it is able to bring down only updated files rather than being forced to always download complete packages.

[Read More]

Just over 7 minutes to install a text-oriented, headless server OpenSolaris system from scratch?

This is the next installment in my effort to establish an efficient development and test environment for experimental OpenSolaris base installation profiles. In this entry I address how to hook a heavily reduced form of the OpenSolaris development package repository into an Automated Installer environment to produce text-oriented, headless server installs of OpenSolaris.

[Read More]

In this post I explain how I've used the pkg(5) system's content retrieval client, pkgrecv(1), to establish a reduced form of the OpenSolaris development package repository on a local system for the purpose of optimizing my Automated Installer (AI)-based installations.

In my recent post on Initial Experiments with Base Install Profiles, I described how I am using the OpenSolaris Automated Installer and VirtualBox to experiment with heavily reduced or minimized installations of OpenSolaris.  Use of a local copy of a repository will dramatically increase the speed of installing these heavily reduced installations.

Use of a local repository also enables me to begin experimenting with publishing and using an experimental base installation profile metapackage.  Use of such a metapackage will greatly simplify the AI manifest by representing all of the base install packages in a single metapackage.  Once a developer includes the base metapackage reference in his AI manifest, all he needs to do is add references to other packages to produce his custom installation profile based on the experimental base profile.[Read More]

This post is a follow-on to the entry describing how to set up an OpenSolaris Automated Installer (AI) environment using the JeOS prototype and VirtualBox.  In this entry I describe how to use a custom AI manifest representing an experimental base installation profile to install a greatly reduced or minimized form of OpenSolaris in a VirtualBox virtual machine - just ~90 or so packages and an installed footprint of ~570 MB. The resulting installations are text-oriented, headless server installations containing only enough features to enable developers to add more packages and customize the environment to suit their specific needs.

A slightly modified form of the AI manifest can be used to realize similarly reduced installations in other system virtualization and bare metal environments.

Experimenting with base installation profiles is part of the OpenSolaris JeOS project's effort to help define a formal base installation profile for OpenSolaris.

[Read More]

This entry addresses how to use the OpenSolaris JeOS prototype and VirtualBox as the basis for deploying an Automated Installer (AI) environment.  A follow-on entry addresses the use of a custom AI manifest to install a greatly reduced form of OpenSolaris using this environment.

[Read More]