| FigDiskAssembly |
UserPreferences |
| SEED Wiki | FrontPage | RecentChanges | TitleIndex | WordIndex | SiteNavigation | HelpContents |
In this document I will discuss what I'm calling the assembly of a FIG disk: the creation of a working SEED system from its component parts:
The perl script 'assemble_disk' will create the basic structure of a FIG disk, including the code release. The Data directory must be created by the user (of assemble_disk; we don't view assemble_disk as the primary mechanism for end-users to obtain a FIG disk).
One common use of assemble_disk is to create a DVD release of a SEED installation.
This procedure would entail the following.
make_transferable_chunks dataDir chunksDir
This will create in chunksDir a set of 100M files containing tarred and gzipped data
that can be resurrected to create a data directory on the target computer.
These are divided up and written to the DVD set.
We will now discuss each of the scripts that comprise the installation suite.
Before any of these are executed, the contents of the disk set (or the FTP download site) must be copied to a scratch location on a disk that has enough space to hold them (it will likely require 10-20G of space above what is required for the SEED install itself).
The following files will be found on the distribution media (DVD set or FTP directory):
The InstallSEED script is the toplevel installation script for a SEED release. This script takes a single argument, the directory into which the SEED will be installed. It is executed from the directory containing the files coped from the disk set.
It performs the following tasks:
This script reads a checksums file from a directory containing data files created by make_transferable_chunks. It takes a single argument, the directory containing the data files. If a file is missing, or if it has an invalid checksum, or if there is some error in the extract, the script will exit with a nonzero return code.