<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.78 [en] (X11; U; Linux 2.4.9-13 i686) [Netscape]">
<link rel="stylesheet" type="text/css" href="../styles/style2.css">
</head>
<body>
Below the steps are described that need to be taken to set up the ASTRO-WISE
Environment database, fileserver and parallel processing parts.
It is assumed that the pipeline is already installed according to the
<a href="http://cvs.astro-wise.org/cgi-bin/viewcvs/cgi/viewcvs.cgi/opipe/README?rev=HEAD&content-type=text/vnd.viewcvs-markup">README</a>
(see also <a href="cvs.html">CVS howto</a>) and that the environment can be
started with the <tt>awe</tt> command.
<ol>
<li>Install the pipeline if you have not already done so and create a databae
account.
</li>
<li>In your <tt>~/.awe/Environment.cfg</tt> set
<pre>database_engine : oracle_9i
database_name : <tnsname of your database here>
database_user : <name of the database account you created previously>
</pre>
Then start <tt>awe</tt> and do <pre>import astro.database.DBMain</pre>
You should get a prompt asking for your database password.
If you are able to connect you can continue with the next step.
</li>
<li>Create the <tt>AWOPER</tt> database account and an <tt>AWUSER</tt> role
<pre>awe opipe/Toolbox/dbawoper.py</pre>
</li>
<li>Create the shared <tt>AWOPER</tt> schema for all database users using
<pre>awe opipe/Toolbox/dbimportall.py</pre>
</li>
<li>Grant <tt>SELECT/INSERT/UPDATE</tt> permission on the <tt>AWOPER</tt>
schema to the <tt>AWUSER</tt> role.
<pre>awe opipe/Toolbox/dbgrants.py</pre>
If new persistent classes are defined you should run this script as well to
make sure that users have access to the new tables.
</li>
<li>Add a new user to the database. The username should consist of the first
initial followed by the completed surname of the user.
<pre>awe opipe/Toolbox/dbnewuser.py mjackson</pre>
Note that this will create a database account with <tt>AW</tt> as a prefix.
In the above example the database account would be called <tt>AWMJACKSON</tt>.
</li>
<li>On the machine where the FITS images should be stored go to the directory
where you want to store them and do
<pre>mkdir cache
awe opipe/parallel/dataserver_server.py <hostname> <port></pre>
The cache directory is the placed where new files will be stored.
</li>
<li>In your <tt>~/.awe/Environment.cfg</tt> add the following lines
<pre>data_server : <hostname_you_have_supplied_to_dataserver_server.py>
data_port : <port_you_have_supplied_to_dataserver_server.py></pre>
</li>
<li>For the parallel processing you have to define the individual nodes on which
you want to run the parallel package.
In your <tt>~/.awe/Environment.cfg</tt> set the following variables (the values are examples):
<pre>
cluster_nodes : virgo01,virgo02,virgo03,virgo04,virgo05,virgo06,virgo07,virgo08
cluster_ports : 9001,9002,9003,9004,9005,9006,9007,9008
cluster_wdirs : /tmp,/tmp,/tmp,/tmp,/tmp,/tmp,/tmp,/tmp
cluster_rsh : ssh -n
cluster_port : 9000
cluster_httpport : 8999
cluster_passwordfile : /Users/users/kgb/.awe/Cluster.pw
cluster_username : kgb
cluster_password : secret
</pre>
The variable <tt>cluster_nodes</tt> defines the different nodes,
<tt>cluster_ports</tt> the ports on which the cluster software is
communicating with the cluster_scheduler and <tt>cluster_wdirs</tt> the
working directories on the diferetn nodes. Note that each node should
have a different working directory and that the path to the opipe
software tree should be the same on all nodes. <br>
The variable <tt>cluster_rsh</tt> defines the command which will be used
to logon on to the individual nodes without user intervention. <br>
<tt>cluster_port</tt> defines the port for the cluster_manager and
<tt>cluster_httpport</tt> the port to get oniine information about the
cluster_manager via a web-browser. <br>
The username/password combination can be generated with the programme
<tt>dbmmanagein</tt> in directory <tt>parallel</tt> of the opipe source
tree. Example of use:
<pre> python /bee1/users/kgb/opipe/parallel/dbmmanage.py
/Users/users/kgb/.awe/Cluster.pw add kgb </pre>
Testing the parallel software can be done with the programme
<tt>testtask.py</tt> in directory <tt>parallel</tt>. </li>
<li>To ingest data use the opipe/Toolbox/ingest.py script.
</li>
</ol>