Hi,
I tried to use the cron block and found some difficulties. Before I
forget it, I'm working with cocoon 2.1.6, java SDK 1.4.2, tomcat 5.0.28
and linux (debian).
Activating the cron block with standard values in the cocoon.xconf
worked fine.
But I wanted to save the cron jobs in a database, so I made one change
in the configuration and was using the JobStore 'tx':
<store type="tx"
delegate="org.quartz.impl.jdbcjobstore.HSQLDBDelegate">
<datasource provider="jndi">jdbc/personnel</datasource>
</store>
(By the way, the only documentation found here was:'tx' and 'cmt' stores
require datasource configuration. Attributes delegate and provider are
optional. Not very helpfull, and which database structure do I have to
use? In the end I found this information on the quatz-scheduler
homepage, but it would be nice to bundle this with the cron block)
First I hoped I could use the datasource defined in the cocoon.xconf,
but I didn't find a solution.
So I had to configure the same datasource again, this time for use as
JNDI Datasource.
I was modifying the server.xml and the cocoon/WEB-INF/web.xml, but I
always got the same error:
org.quartz.SchedulerConfigException: Failure occured during job
recovery. [See nested exception: org.quartz.JobPersistenceException:
Failed to obtain DB connection from data source 'jndi:jdbc/personnel':
java.sql.SQLException: Could not retrieve datasource via JNDI url
'jdbc/personnel' javax.naming.NameNotFoundException: Name jdbc is not
bound in this Context [See nested exception: java.sql.SQLException:
Could not retrieve datasource via JNDI url 'jdbc/personnel'
javax.naming.NameNotFoundException: Name jdbc is not bound in this
Context]]
Then I got the sourcecode of the quartz scheduler and found where I was
running into problems:
src/java/main/org/quartz/utils/JNDIConnectionProvider.java
the function init():
private void init() {
if (!isAlwaysLookup()) {
try {
Context ctx = null;
if (props != null) ctx = new InitialContext(props);
else
ctx = new InitialContext();
datasource = (DataSource) ctx.lookup(url);
} catch (Exception e) {
getLog().error(
"Error looking up datasource: " +
e.getMessage(), e);
}
}
}
Changing it into:
private void init() {
if (!isAlwaysLookup()) {
try {
Context ctx = null;
if (props != null) ctx = new InitialContext(props);
else
ctx = new InitialContext();
Context envCtx = (Context) ctx.lookup("java:comp/env");
datasource = (DataSource) envCtx.lookup(url);
} catch (Exception e) {
getLog().error(
"Error looking up datasource: " +
e.getMessage(), e);
}
}
}
and compiling and replacing it with the quartz jar in the cocoon package
seems to solve the problem.
I don't know enough about JNDI to tell if this is a bug or if this is
just a house made problem. Can anyone tell me?
By the way, using the hsqldb database which is included in cocoon with
this aproach it's not possible, I got always a JNDI/SQL Error because
the database wasn't up, and because of this (early) error, the database
wasn't coming up. I had to use a standalone database.
And when I'm right here, the commentar to the standard values must be
fixed, too, some are wrong, e.g.:
<!--+
| How big should the queue be.
| Defaults to unlimited size (<0 == default)
+-->
<queue-size>-1</queue-size>
<!--+
| The maximum size of the pool.
| Defaults to Integer.MAX_VALUE (<0 == default)
+-->
<max-pool-size>-1</max-pool-size>
At least one of them isn't working with a value of 0.
