File	Date	Author	Commit
contrib	2012-03-30	Simon G	[cec320] ...
src	2013-04-25	Simon G	[1d62a7] removed some log messages
.gitignore	2012-03-29	Simon G	[b992ab] initial commit: Job Server Version 1.0
LICENSE	2013-04-25	Simon G	[27ac88] added support for specific solver binary id for...
Makefile	2012-03-29	Simon G	[c22652] allow static compilation, updated help, updated...
README	2012-03-29	Simon G	[c22652] allow static compilation, updated help, updated...

Read Me

Job Server 1.0
==============

Job Server for EDACC (Experiment Design and Administration for Computer Clusters).
The Job Server is optional but can improve overall performance of job scheduling.
See http://sourceforge.net/projects/edacc/ for the EDACC project.

Installation
============

Compile the Job Server binary by using make. This will create a binary called jobserver
in the bin directory.

Starting Job Server on startup
===============================

In contrib/etc/init/edacc_jobserver.conf is an example upstart script given. If your
Linux Distribution supports upstart, you can use that script to start the Job Server
automatically.

Create a /var/log/edacc_jobserver/ directory for the log files.
Create a /etc/edacc directory for the configuration file.
Copy the configuration file in contrib/etc/edacc/jobserver.conf to /etc/edacc
Change the configuration file by your needs.
Copy the jobserver binary to /usr/bin/.
Copy the upstart script in contrib/etc/init to your upstart init directory (/etc/init).
Change the username in the upstart init script to an user (unprivileged) that exists.

Use service edacc_jobserver start to start the jobserver.


Usage
=====

The Job Server needs a configuration file. An example for a configuration file is given
in contrib/etc/edacc/jobserver.conf.

Command Line Options are:
  --logdir=<log dir>		use the <log dir> to log files. The Job Server will create a new
  							log file called jobserver.log on every start. If jobserver.log
  							exists it will be moved to jobserver_<number>.log.
  --config=<config file>	the configuration file for the Job Server.
  --verbosity=<number>		the verbosity for logging.
  --user=<user name>		change to user <user name> on startup. It is recommended that you
  							change to an unprivileged user when starting the Job Server as
  							the root user.
  --group=<group name>		change to group <group name> on startup
  --help					prints the help
  

Client Setup
============

In order to let a client connect to the Job Server you have to specify the Job Server in the
client's configuration by adding those two lines:
jobserver_host = <host>
jobserver_port = <port>
The client will then use the Job Server to retrieve jobs.