Variable Name | Meaning |
---|---|
PREDATOR_LAIR | A path to the directory in which the distribution was downloaded and unpacked. By convention, this directory is usually named PREDATOR. At Cornell, this directory is /usr/u/praveen/PREDATOR. |
XXX_INSTALLROOT | A path to the root of the installation (not the distribution!). By convention, this should point to a directory called predator somewhere. Make sure that you create such a directory. The default value is $HOME/predator. |
XXXSHOREDEFDEVPATH | The desired location of your "default disk". This is the Unix file that is used by PREDATOR ADTs as a storage volume. For instance, images and audio files are stored in this file. The default value is $XXXROOT/defdisk. Try to place your "default disk" on a local file system. |
XXXSHOREDEFDEVSIZE | The size of your defdisk, in kilobytes. A good starting size is 10000. |
XXXSHORERELDISKPATH | The desired location of your relational data "disk". This is the Unix file that is used by the PREDATOR Relation-ADT for storage. A good starting value is $XXXROOT/reldisk. Try to place your "relational disk" on a local file system. |
XXXSHORERELDISKSIZE | The size of your reldisk, in kilobytes. A good starting size is 10000. |
XXXSHORELOGDIR | The directory in which SHORE will store its log files. This directory MUST exist and be writeable. Try to place your log files on a local disk. You MUST define this variable. |
XXXTMPDATADIR | The directory which will serve as the "buffer" between the PREDATOR and web servers. This directory MUST exist and be writeable. The database server will write out temporary files such as images and query processing information files to this directory, and the webserver will link to it to retrieve them. You MUST define this variable. |
XXXSHOREDISKRW | The location of SHORE's diskrw program, which performs asynchronous disk accesses. By default, this has the value $PREDATOR_LAIR/shore/apr97/bin/diskrw. Change this only if you are using an independent installation of SHORE. |
XXX_SERVER_PORT | A port number (preferably above 5000) where the regression tests can safely start a server. This is meaningful only for the regression tests, not for normal execution. |
PATH | In order for the server to be able to use Java classes (such as the one it uses to fetch data through URLs), it must have access to your JDK installation. Add a path to your JDK bin directory to your PATH variable. |
Note: The string "XXX" in this document really means "XXX". Do not subsitute with other strings.
mkdir $XXX_INSTALLROOT cd $XXX_INSTALLROOT ln -s $XXX_RCSROOT/RCS .Now if you type ls, you should see the link to the RCS directory.
co RCS/*This will check out a copy of the Makefile and copyright notices. The Makefile can be used to complete the installation of PREDATOR. Now type
make codeCompilation is a lengthy process (approximately 30 minutes on a SparcStation 20). You will receive an e-mail informing you if the compilation was successful.
If you have trouble with the installation, it is probably because one of the necessary utilities was not available or was improperly specified. Try to fix the problem from the error message, and repeat "make code". Otherwise, send mail with the typescript file to predator-support@cs.cornell.edu.