SAS installation

Introduction

Each new release of SAS is provided in a single tgz archive with all its components, including eventually some additional patches, and a shell script used to install it (install.sh). The tgz archives have the generic names sas_N.N.N-OS.tgz, where

• N.N.N is the SAS version number, e.g. 11.0.0
• OS is an acronym of the Operating System, Version and Kernel word length where such release was built, e.g. Darwin-10.3.0-32.

Examples :

• sas_11.0.0-Fedora12-32.tgz contains the SAS 11.0.0 release built on Fedora 12 32 bit kernel.
• sas_11.0.0-Darwin-9.8.0-32.tgz contains the SAS 11.0.0 release built on a Mac OS X running Darwin 9.8.0 (Mac OS X 10.5.8 Leopard), 32 bit kernel.

For a complete list of Operating Systems and Versions used to build the current release of SAS, please look here. In this page you will find as well indications on what Operating Systems and Versions you might run the builds provided.

All SAS installations are binary. We do not provide support on building SAS from source code. However, all the SAS source code is available for each release and can be downloaded for reference.

When installing SAS, do not forget to consult the page on the external tools required to work together with it. Missing any of these tools would not allow you to work with SAS properly.

Please take into account that for running SAS you will need also the calibration data contained in the so-called Current Calibration Files (CCF). Go to the CCF pages for learning how to download and install them.

Installation procedure common to all Operating Systems

First download the most appropriate tgz archive for your operating system and put it into a directory of your choice, for example let this be /top_dir.

Next, unpack the tgz archive with the command

tar zxf sas_N.N.N-OS.tgz

and inmediately after, execute the install.sh script as follows

./install.sh

and let it run until completion.

The installation shell script will check for the presence of all required SAS software components according to an installation manifest named sas_N.N.N_install_manifest-OS.txt, install all in turn and configure them for inmediate use.

The installation script creates below /top_dir a directory named xmmsas_20110223_1801 (xmmsas_20110223_1803 on Mac OS X), which contains the whole SAS software and launches configure_install that will create in the same directory two scripts named setsas.sh and setsas.csh. These scripts can be used later on to intialize properly the recently installed SAS.

The configure_install shell script makes several sanity checks about the installation and informs the user about them. The most important of these checkings is the search for the existence of perl in /usr/local/bin, either as a real binary or as a soft link to a binary placed in another location, typically in /usr/bin. Should your installation miss perl in the expected location, please proceed to correct the situation and re-run the configure_install shell script again.

Initialization

Depending on your shell, source either setsas.sh or setsas.csh, as follows:

. ./setsas.sh (sh/bash/ksh)

source ./setsas.csh (csh/tcsh)

The previous commands should be replaced by

. /top_dir/xmmsas_20110223_1801/setsas.sh (sh/bash/ksh)

source /top_dir/xmmsas_20110223_1801/setsas.csh (csh/tcsh)

when invoked from any other place of your system.

The setsas.[c]sh scripts take care of all the necessary details to initialize properly the environment to run SAS in your operating system, except for the SAS_CCFPATH, the SAS_ODF and the SAS_CCF environment variables. These must be defined by you explicitly to tell SAS where to find the XMM-Newton observation data you want to analyze and the calibration files needed to do it.

Since SAS 10.0.0, it is mandatory to initialize the HEASOFT/HEADAS software before initializing the SAS. Otherwise the initialization of SAS issues an error message and it is aborted.

For more details consult the documentation links.