jmxinvokerjob.java

定时器开源项目, 相对于 jcrontab, Quartz 算是更完整的一个项目, 随著开发的版本上来, 他已经脱离只是写在程序里面的计时器, 在指定的时间或区间, 处理所指定的事件. 也加入了 se

package org.quartz.jobs.ee.jmx;


import java.util.LinkedList;
import java.util.StringTokenizer;

import javax.management.MBeanServer;
import javax.management.MBeanServerFactory;
import javax.management.ObjectName;

import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.quartz.Job;
import org.quartz.JobExecutionContext;
import org.quartz.JobExecutionException;


/**
 * Generic JMX invoker Job.  It supports any number or type of parameters
 * to the JMX bean.<p>
 * 
 * The required parameters are as follows (case doesn't matter):<p>
 * <dl>
 * <dt><strong>JMX_OBJECTNAME</strong>
 * <dd>This is the fully qualifed name of the object (ie in JBoss to lookup
 * the log4j jmx bean you would specify "jboss.system:type=Log4jService,service=Logging"
 * <dt><strong>JMX_METHOD</strong>
 * <dd>This is the method to invoke on the specified JMX Bean. (ie in JBoss to
 * change the log level you would specify "setLoggerLevel"
 * <dt><strong>JMX_PARAMDEFS</strong>
 * <dd>This is a definition of the parameters to be passed to the specified method
 * and their corresponding java types.  Each parameter definition is comma seperated
 * and has the following parts: <type>:<name>.  Type is the java type for the parameter.  
 * The following types are supported:<p>
 * <b>i</b> - is for int<p>
 * <b>l</b> - is for long<p>
 * <b>f</b> - is for float<p>
 * <b>d</b> - is for double<p>
 * <b>s</b> - is for String<p>
 * <b>b</b> - is for boolean<p>
 * For ilfdb use lower for native type and upper for object wrapper. The name portion
 * of the definition is the name of the parameter holding the string value. (ie
 * s:fname,s:lname would require 2 parameters of the name fname and lname and
 * would be passed in that order to the method.
 * 
 * @author James Nelson ([e-mail]) -- Provident Solutions LLC
 * 
 */
public class JMXInvokerJob implements Job {

	public void execute(JobExecutionContext context) throws JobExecutionException {
		try {
			Object[] params=null;
			String[] types=null;
			String objName = null;
			String objMethod = null;
			String[] keys = context.getJobDetail().getJobDataMap().getKeys();
			for (int i = 0; i < keys.length; i++) {
				String value = context.getJobDetail().getJobDataMap().getString(keys[i]);
				if ("JMX_OBJECTNAME".equalsIgnoreCase(keys[i])) {
					objName = value;
				} else if ("JMX_METHOD".equalsIgnoreCase(keys[i])) {
					objMethod = value;
				} else if("JMX_PARAMDEFS".equalsIgnoreCase(keys[i])) {
					String[] paramdefs=split(value, ",");
					params=new Object[paramdefs.length];
					types=new String[paramdefs.length];
					for(int k=0;k<paramdefs.length;k++) {
						String parts[]=  split(paramdefs[k], ":");
						if(parts.length<2)
							throw new Exception("Invalid parameter definition: required parts missing "+paramdefs[k]);
						switch(parts[0].charAt(0)) {
							case 'i':
								params[k]=new Integer(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Integer.TYPE.getName();
								break;
							case 'I':
								params[k]=new Integer(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Integer.class.getName();
								break;
							case 'l':
								params[k]=new Long(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Long.TYPE.getName();
								break;
							case 'L':
								params[k]=new Long(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Long.class.getName();
								break;
							case 'f':
								params[k]=new Float(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Float.TYPE.getName();
								break;
							case 'F':
								params[k]=new Float(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Float.class.getName();
								break;
							case 'd':
								params[k]=new Double(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Double.TYPE.getName();
								break;
							case 'D':
								params[k]=new Double(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Double.class.getName();
								break;
							case 's':
								params[k]=new String(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=String.class.getName();
								break;
							case 'b':
								params[k]=new Boolean(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Boolean.TYPE.getName();
								break;
							case 'B':
								params[k]=new Boolean(context.getJobDetail().getJobDataMap().getString(parts[1]));
								types[k]=Boolean.class.getName();
								break;
						}
					}
				}
			}
			
			if(objName==null || objMethod==null) 
				throw new Exception("Required parameters missing");
			
			invoke(objName, objMethod, params, types);
		} catch (Exception e) {
			String m = "Caught a " + e.getClass().getName() + " exception : " + e.getMessage();
			getLog().error(m, e);
			throw new JobExecutionException(m, e, false);
		}
	}
  
  private String[] split(String str, String splitStr) // Same as String.split(.) in JDK 1.4
  {
    LinkedList l = new LinkedList();
    
    StringTokenizer strTok = new StringTokenizer(str, splitStr);
    while(strTok.hasMoreTokens()) {
      String tok = strTok.nextToken();
      l.add(tok);
    }
    
    return (String[])l.toArray(new String[l.size()]);
  }

	private void invoke(String objectName, String method, Object[] params, String[] types) throws Exception {
		MBeanServer server = (MBeanServer) MBeanServerFactory.findMBeanServer(null).get(0);
		ObjectName mbean = new ObjectName(objectName);

		if (server == null)
			throw new Exception("Can't find mbean server");

		getLog().info("invoking " + method);
		server.invoke(mbean, method, params, types);
	}

	private static Log getLog() {
		return LogFactory.getLog(JMXInvokerJob.class);
	}

}