Defining Your Local SNODB Configuration

  The database setup proceeds through a single perl script:

$\sim$/3_06_02/scripts/snodb.perl

NOTE: If this does not work please edit the file and replace the top line with the location of your local node's perl executable. Prior to actually running snodb.perl you need to edit the file:

$\sim$/3_06_02/scripts/snodb.config

to specify the location of various directories (see footnote below for VMS-specific information) and various snodb initializations.

One of the fields (labelled ``official'' or ``test'') is used to define whether an official SNO database is set up for general use or if a test database is set up for personal use (for testing data before it is input into the official database). Leave the field as ``official.''

The final step is to set up some environmental variables:

UNIX:

set the environmental variable CDSERV which tells HEPDB where your server resides (i.e., under the parent directory with the name you have chosen for your server). The default is /hepdb/cdsno and it is set by the snodb.perl script as needed. However, other users who wish to run their own personal server (for example, with test database files) will need to edit their .login file and define the variable CDSERV to point to their personal server directory. For example, in csh:

setenv CDSERV /users/robk/hepdb/cdtest

The environmental variable SNODB_VER needs to be set to the name of the SNODB version, the variable SNODB_ROOT needs to be set to the name of the directory in which the ``3_06_02/scripts...'' directory structure resides, and the variable SNO_CODE needs to be set to the directory containing the latest development release of SNOMAN. For example, in csh:

setenv SNODB_ROOT /users/cdsno
setenv SNODB_VER 3_06_02
setenv SNO_CODE /users/snoman/dev

VMS:

both a global symbol and a logical name called CDSERV have to be defined as described in the UNIX paragraph above. Logical names SNODB_ROOT, SNODB_VER and SNO_CODE also have to be defined as described above. For example:

$ define CDSERV dkb$200:[hepdb.cdsno]

$ cdserv == ``dkb$200:[hepdb.cdsno]''

$ define SNODB_ROOT dkb$200:[cdsno]

$ define SNODB_VER 3_06_02

$ define SNO_CODE dkb$200:[snoman]

It is important that the global symbol CDSERV and the logical names SNODB_ROOT and SNODB_VER be accessible to batch processes, so be careful that you don't define them only under a IF F$MODE() .EQS. "INTERACTIVE" THEN... -type statement in your LOGIN.COM file.

A global symbol pointing to the EXTRACT.EXE image needs to be defined. For example:

$extract:==$dkb$200:[snoman.tools.exe]extract.exe

The name of the batch queue in which the HEPDB database server will run needs to be defined using the logical name CDSERV$BATCH_QUEUE. The queue should allow infinite maximum CPU time, and the owner of the server job should have permission to start and stop the queue. A high priority for the queue is not necessary.

Most of the variables defined above are only required by the account from which the database is maintained (i.e. the cdsno account). Others who simply want read or write access to database files only need to set up the CDSERV environmental variable on UNIX systems and the CDSERV global symbol on VMS systems as defined above.

The setup script automatically loads in the compiling and linking flags defined for your platform in the SNOMAN file $SNO_TOOLS/get_platform_variables.scr. For this to work you must have edited the contents of that file as needed and defined SNO_TOOLS and SNO_PLATFORM (see the SNOMAN installation procedure).

You are now ready to run snodb.perl. To see a list of the snodb.perl command options, type the command:

UNIX:

$\sim$/3_06_02/scripts/snodb.perl -help

VMS:

perl [cdsno.3_06_02.scripts]snodb.perl -help

Some tedious typing can be avoided if the following is done:

UNIX:

put the /3_06_02/scripts directory in your path, then simply type:
snodb.perl -help

VMS:

define a symbol to run the script:
snodb_cmd=="''perl' dkb$200:[cdsno.3_06_02.scripts]snodb.perl"
snodb_cmd -help

Note that in what follows command options are given one at a time, but that they can be entered simultaneously if you wish, keeping in mind that the order may matter. Also note that, on UNIX systems, if /usr/bin/perl does not exist (either as an executable or as a link to one) then you will either have to edit the first line in snodb.perl to point to the location of the executable, or type perl before the script name like on VMS systems.