Page maintained by
Configuring and running the DataBase Handler.
After installing ISDAT you have to configure the server configuration file isdat.server in the directory
For each project (=server module) there are a number of entries that must be set in
order for the server to work correctly. Their names and number varies between
projects. Typically they will tell the server in witch
directorys to look for the various datafiles.
Syntax in isdat.server
The syntax of isdat.server is quite flexible. isdat.server is designed to be portable between computers. It therfore supports different configurations for different computers in the same file.
The main concept of the file is to write the name of the entry on the left, then directly a colon as an separator, thereafter comes the actual value.
A entry name has the form
It is possible to use a wildcard (*) instead of the computer name. This will
match all computers but only after any direct matches. There is a project entry
called common in this file, it stands for configurations entrys in the server
that are same for all server modules.
A data field can have any form but the most common is a directory path. You
can use shell variables such as ISDAT_HOME in these paths. They will be passed
on to a shell to be expanded. For some modules is it possible provide several
directorys bye separating them with a ´:´.
Finally, a comment is an row that starts with a # sign.
The Cluster example.
The following rows is taken out of a isdat.server file. They tell the wec module were to look for its datafiles, were to write log files and so on.
# stuff for the cluster satellite iroc.cluster.rdm: /data/cluster/RDM:/cdrom iroc.cluster.dds: /data/cluster/DDS iroc.cluster.index: /data/cluster/index iroc.cluster.burst: /data/cluster/burst iroc.cluster.cal: /data/cluster/cal iroc.cluster.log: /data/cluster/log plutt.cluster.rdm: /data/cluster/RDM plutt.cluster.dds: /data/cluster/DDS plutt.cluster.index: /data/cluster/index plutt.cluster.burst: /data/cluster/burst plutt.cluster.cal: /data/cluster/cal plutt.cluster.log: /data/cluster/log efw.cluster.rdm: /data/cluster/RDM:/cdrom efw.cluster.dds: /data/cluster/DDS:/data/cluster/DDS_AL efw.cluster.index: /data/cluster/index efw.cluster.burst: /data/cluster/burst efw.cluster.cal: /data/cluster/cal efw.cluster.log: /data/cluster/log #*.cluster.rdm: $HOME/data/cluster/RDM:/cdrom *.cluster.rdm: /cdrom *.cluster.dds: $HOME/data/cluster/DDS *.cluster.index: $HOME/data/cluster/index *.cluster.burst: $HOME/data/cluster/burst *.cluster.cal: $HOME/data/cluster/cal *.cluster.log: $HOME/data/cluster/log *.cluster.DSNdata: /mnt/cdrom/cluster2First you will notice that there are several sections for different computers. If any of the first ones fails to match, the module will use the last section with wildcards for the computer name and $HOME/data/ as root for the data directorys. Secondly you will see that some of the rows sets its entry to a list of directorys. A search path just like your shell PATH environment variable.
To know what entrys are avaliable for a module and what to set them to refer to the documentation for that module (The isdat.server distributed with the source code contains defaults for most modules).
Starting the ISDAT server
Time to start the database handler. First make sure all necessary environment variables are set. Now you can start the server by typing:
where the number can be any number, but in order for your clients to contact the
database it must correspond to the value set in
ISDAT_DATABASE. Also it should not conflict with the number of
any another dbh running on your computer.