Monday, June 6, 2011

Implementing a Plugin in Quartz.Net 2.0

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

Today we will implement a plugin for Quartz.Net 2.0. If you’re looking for a way to implement a plugin for version 1.0 of Quartz.Net, take a look at this post. Most of the steps are the same, but there are some API changes, so the plugin isn’t exactly the same, as you will see.
What are plugins used for in Quartz.Net 2.0? Well, plugins come in handy if there is something you want to do whenever the scheduler is started or stopped. Some examples of handy plugins are:
  • plugins to load jobs into a RAM jobstore upon scheduler startup (this one is available out of the box)
  • plugins to register listeners (listeners must be registered with the scheduler upon startup since they are not persisted)
In order to implement a listener in Quartz.Net 2.0 there are several steps we need to take. First we must implement the ISchedulerPlugin interface and then we must configure the scheduler to load our plugin. Let’s get started.

Implementing the ISchedulerPlugin Interface

In order to create your own plugin, you have to implement the ISchedulerPlugin interface. Here is the ISchedulerPlugin interface definition:
public interface ISchedulerPlugin
 {
    void Initialize(string pluginName, IScheduler sched);
    void Start();
    void Shutdown();
 }



As you can see it’s not a very complicated interface. Let’s take some time to look at these methods in a little more detail.


Initialize


The Initialize method is called before the scheduler is started and it gives your plugin a chance to initialize itself. Keep in mind that at this point the scheduler itself is not running nor fully initialized. Here is what the code comments have to say:



Called during creation of the IScheduler in order to give the ISchedulerPlugin a chance to Initialize. At this point, the Scheduler's IJobStore is not yet (initialized?). If you need direct access (to) your plugin, you can have it explicitly put a reference to itself in the IScheduler’s SchedulerContext (in the) Initialize(string, IScheduler) method.


Start


The Start method is probably the method where most of the logic for your plugin will go. I will let the method’s comments describe the method’s purpose:


Called when the associated IScheduler is started, in order to let the plugin know it can now make calls into the scheduler if it needs to.


Shutdown


The shutdown method is there to allow you to do any cleanup that might be necessary for your plugin. The code comments are as follows:



Called in order to inform the ISchedulerPlugin that it should free up all of it's resources because the scheduler is shutting down.


Sample Implementation



Now we will go ahead and actually implement the plugin. In our sample implementation all that our plugin will do is to register a job listener. We are following the same approach we took while discussing how to implement a plugin in version 1.0 of Quartz.Net. This is intentional, to allow you see what the differences are between the two versions.


The implementation of our plugin is still very simple, albeit a little different from the one on the previous post:




public class PluginExample : ISchedulerPlugin
 {
    public void Initialize(string pluginName, IScheduler sched)
    {
        sched.ListenerManager.AddJobListener(new JobListenerExample(), EverythingMatcher<JobKey>.AllJobs());
    }

    public void Shutdown()
    {
        //Do Nothing
    }

    public void Start()
    {
        //Do Nothing
    }
 }



Configuring the Scheduler to Load the Plugin



Now that we have created our plugin, the next step is to tell the scheduler to load it. This is done by adding this line to our properties file :

quartz.plugin.myplugin.type = Examples.PluginExample, Examples



The plugin property needs to be structured like this:


quartz.plugin.{name}.type = {type name}, {assembly name}

Let’s discuss the components of the property. First, the quartz.plugin part of the property tells the scheduler that this is a plugin that needs to be loaded. The {name} part is the name that you want to give to your plug-in, and is the name that is passed in to the Initialize method above. The value of the property {type name}, {assembly name} tells the scheduler the Type of the plug-in, so that it can be loaded into memory. {type name} is the full name of your plugin, which in our case would be Examples.PluginExample. {assembly name} is the name of your assembly file, minus the .dll extension. Here’s another example of a plugin configuration property, lifted from the Quartz.Net default configuration file:


quartz.plugin.xml.type = Quartz.Plugin.Xml.XMLSchedulingDataProcessorPlugin, Quartz


Final Step



Now that your plugin has been added to the configuration, all you need to do is start the scheduler and check the log for errors. Remember that the dll that contains your plugin needs to be in the same folder as the Quartz.Net binaries so that it can be found by the .Net framework and loaded into memory.

Tuesday, March 15, 2011

New in Quartz.Net 2.0–New Job File Format

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

The xml format for the quartz_jobs.xml file has changed in Quartz.Net 2.0. This is a breaking change, so you won’t be able to use your existing jobs file with the new version of Quartz.Net without updating it to the 2.0 format. If you’re interested in the details of what can be configured in this file, I would recommend looking at the xsd file that defines the schema for the file. The xsd file that defines the new file format can be downloaded directly from the source code repository, here.
Today I’m going to provide a sample quartz_jobs.xml file in the new format and walk you through the creation of the file. Alternatively, you can take a look at the default file (that will be) provided with the Quartz.Net 2.0 distribution, here.
The new file format starts off with the following root element:
<job-scheduling-data xmlns="http://quartznet.sourceforge.net/JobSchedulingData" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="2.0">

Right after the root node, our configuration file has two elements, which we will describe in detail below:
  1. The <processing-directives> element

  2. The <schedule> element

The <processing-directives> Element

The processing directives element contains two elements. The first element, <pre-processing-commands> is used to run commands prior to adding jobs and triggers to the scheduler. I’m not going to spend much time going over these in detail, but here is a list of the commands available to you:

  • delete the jobs in a group
  • delete the triggers in a group
  • delete a job
  • delete a trigger
If you find yourself in need of help with these elements, leave a comment and I will write a follow-up post.

The second element, <processing-directives>, lets you specify how the xml file is to be processed. Here you will find a familiar configuration setting (from Quartz.net 1.0), the <overwrite-existing-data> element. Setting this element’s value to true will replace any jobs currently scheduled with the new schedule given in the xml file (true is the default setting!). The <ignore-duplicates> element is the other element allowed under the <processing-directives> element.

The <schedule> Element

The <schedule> element is where we describe the jobs and triggers that we want to schedule. This section has also changed its format. Now, instead of grouping jobs and triggers under an element, jobs and triggers are all specified at the same level, and are not grouped under the job element, as was the case in version 1.0.

Under the <schedule> element, we can have <job> and <trigger> elements, which are the building blocks that we use to put together our schedule. Let’s take a look at these elements now.

The <job> Element

The <job> element is used to describe the IJob that we want the scheduler to execute. Job details are specified by using the following elements:
  • <name>
  • <group>
  • <description>
  • <job-type>
  • <durable>
  • <recover>
  • <job-data-map>
All of these elements serve the same purpose as they did in version 1 and, except for the job-data-map, they’re all simple types, so I’m not going to go into great detail here either. If you need a little more help with these, leave a comment and I’ll expand on the explanation as needed.

The job-data-map, being a complex element, supports an <entry> element inside. The <entry> element has <key> and <value> elements inside, which describe the job property to be added to the job map.

The <trigger> Element
The <trigger> element is used to describe the trigger that we want to attach to a given job. The xml file loader plugin supports 3 types of triggers:
  • simpleTriggerType
  • cronTriggerType
  • calendarIntervalTriggerType
These three trigger types are based on the abstractTriggerType, which has the following common elements:
  • <name>
  • <group>
  • <description>
  • <job-name>
  • <job-group>
  • <priority>
  • <calendar-name>
  • <job-data-map>
In addition to the elements listed above, each trigger type supports some additional elements. These additional elements are used to set parameters that are not common across all of the trigger types. The table below summarizes each of the trigger types and the elements they support.

Trigger Type Additional Elements
SimpleTrigger <misfire-instruction>,<repeat-count>,<repeat-interval>
CronTrigger <misfire-instruction>,<cron-expression>,<time-zone>
CalendarIntervalTrigger <misfire-instruction>,<repeat-interval>,<repeat-interval-unit>


I’m not going to go into detail here either. This is already a pretty long post, so if anybody needs more information leave me a comment and I will write a follow-up post.

Putting It All Together

At this point we’ve described all of the elements that are necessary to describe a job in the new Quartz.Net 2.0 xml format. To put it all together, here are the contents of a sample quartz_jobs.xml file that you can use as a guide:

<?xml version="1.0" encoding="UTF-8"?>

<!-- This file contains job definitions in schema version 2.0 format -->

<job-scheduling-data xmlns="http://quartznet.sourceforge.net/JobSchedulingData" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="2.0">
  <processing-directives>
    <overwrite-existing-data>true</overwrite-existing-data>
  </processing-directives>
  <schedule>
    <job>
      <name>nativeJobExample</name>
      <group>nativeJobExampleGroup</group>
      <description>Sample job for Quartz Server</description>
      <job-type>Quartz.Job.NativeJob, Quartz</job-type>
      <job-data-map>
        <entry>
          <key>command</key>
          <value>native_job_example.bat</value>
        </entry>
        <entry>
          <key>consumeStreams</key>
          <value>true</value>
        </entry>
      </job-data-map>
    </job>
    <trigger>
      <simple>
        <name>nativeJobExampleSimpleTrigger</name>
        <group>nativeJobExampleSimpleTriggerGroup</group>
        <description>Simple trigger example</description>
        <job-name>nativeJobExample</job-name>
        <job-group>nativeJobExampleGroup</job-group>
        <misfire-instruction>SmartPolicy</misfire-instruction>
        <repeat-count>5</repeat-count>
        <repeat-interval>10000</repeat-interval>
      </simple>
    </trigger>
  </schedule>
</job-scheduling-data>

This is a working example of a quartz_jobs.xml file that schedules a NativeJob using a SimpleTrigger. This example also shows how to configure the <job-data-map> to pass configuration information to the job.

I hope this post helps you in building a job configuration file for Quartz.Net 2.0.

Wednesday, March 9, 2011

Scheduling Jobs Programmatically in Quartz.Net 2.0

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

Scheduling jobs programmatically in Quartz.Net 2.0 is very similar to scheduling a job in Quartz.Net 1.0. The overall process is the same, but the scheduler API has changed for version 2.0, so the code used will be significantly different.
In this post we will only highlight the differences between the two versions, so if you’re not familiar with the process, then read this post first. Then come back to this post to understand what has changed in version 2.0.

What Has Changed

Only a couple of things have changed in Quartz.Net 2.0 when it comes to scheduling a job programmatically: the way you build jobs and triggers.
The new Quartz.Net scheduler API uses more interfaces now. It also provides builder objects to create triggers (ITrigger now) and job details (IJobDetail now). I think the simplest way to highlight the differences is to take the code from my original post and convert it to the new API. Here’s the updated code:
public void ScheduleOneTimeJob(Type jobType, JobDataMap dataMap)
{
  string name = DateTime.Now.ToString("yyyyMMddHHmmss"));
  string group = "OneTimeJobs";
  // create job
  IJobDetail jobDetail = JobBuilder
    .Create()
    .OfType(jobType)
    .WithIdentity(new JobKey(name, group))
    .UsingJobData(dataMap)
    .Build();

  // create trigger
  ITrigger trigger = TriggerBuilder
    .Create()
    .ForJob(jobDetail)
    .WithIdentity(new TriggerKey(name, group));
    .WithSchedule(SimpleScheduleBuilder.Create())
    .StartNow()
    .Build();
  // get reference to scheduler (remote or local) and schedule job
  GetScheduler().ScheduleJob(jobDetail, trigger);
}



As you can see, the changes to the code are pretty significant, but the overall process remains the same. I’d like to provide you with some links to documentation or some other examples, but I’m unaware of any at this point, as Quartz.Net 2.0 has not yet been released.

Scheduling Jobs Programmatically in Quartz.Net 1.0

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

This post is one of those back to basics posts and is inspired by a comment left by Jan on one of the Getting Started posts. Today we’ll take a look at how to schedule a job programmatically in Quartz.Net 1.0. I’ll follow up with a post on how to schedule a job programmatically in Quartz.Net 2.0 as well.

Scheduling Jobs in Quartz.Net

There are two ways you can schedule a job to run on Quartz.Net. The first way is to schedule it using the plugin that reads an xml file containing the job and scheduling information. If you want more information on how to schedule jobs using the xml plugin, take a look at this post.
You can also schedule jobs by interacting directly with the scheduler API. The Quartz.Net scheduler supports remote access via the .Net remoting facility, so this method of scheduling jobs is available to you regardless of whether the scheduler is running as a standalone windows service or whether it is embedded in your application. The remote scheduler does have some limitations, but we will not run into them while scheduling jobs. Let’s take a look at how we can schedule jobs using the scheduler API.

Scheduling Jobs via the Scheduler API

In a nutshell, these are the steps you would follow to schedule a job using the API:
  1. Get a reference to the scheduler object. How you do this will depend on whether you are running a standalone or embedded scheduler.
  2. Create a JobDetail object and set all the necessary properties. At a minimum you must set the Name, the Group and JobType. Please note that because the type of JobType is Type, the dll containing the JobType MUST be available to the code that is scheduling the job AS WELL as to the scheduler.
  3. Create a Trigger object (Trigger itself is an abstract class, so you will actually be creating one of these concrete types: SimpleTrigger, CronTrigger or NthIncludedDayTrigger) and set all the necessary properties. At a minimum you must set the Name, and the Group. The JobGroup and the JobName are also necessary as they are what links the JobDetail you created in the previous step with the Trigger you are creating now. However, you do not need to set these properties right now. Read on to step 4 to understand why.
  4. Call the ScheduleJob method on your scheduler object passing in your Trigger and your Job. If you call this overload of the ScheduleJob method, then the scheduler will link your trigger and your job, so there is no need for you to set the link explicitly.

Code Example

I’d like to close out this post with a code sample that implements the steps outlined above.
public void ScheduleOneTimeJob(Type jobType, JobDataMap dataMap)
{
  string name = DateTime.Now.ToString("yyyyMMddHHmmss"));
  string group = "OneTimeJobs";

  // create job
  JobDetail jobDetail = new JobDetail(name, group, jobType);
  jobDetail.Description = "One time job.";
  jobDetail.JobDataMap = dataMap;
  
  // create trigger
  SimpleTrigger trigger = new SimpleTrigger();
  trigger.Name = name;
  trigger.Group = group;
  trigger.StartTimeUtc = DateTime.UtcNow;
  trigger.RepeatCount = 0;
  trigger.RepeatInterval = TimeSpan.Zero;
 
  // get reference to scheduler (remote or local) and schedule job
  GetScheduler().ScheduleJob(jobDetail, trigger);
}



This sample code demonstrates one way of creating and scheduling a job that runs once, immediately.


In Quartz.Net 2.0 the scheduler API has changed significantly, so we will cover that topic in another post.

Tuesday, March 1, 2011

How Does Quartz.Net Configuration Work?

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

In this post we will explain in great detail how the Quartz.Net scheduler loads its configuration upon startup. Since the scheduler configuration did not change between versions 1.0 and 2.0, the information here is applicable to both Quartz.Net versions.
In Quartz.Net, the StdSchedulerFactory is responsible for configuring the scheduler. When the Quartz.Net scheduler is started, the factory will try to automatically configure a scheduler by looking for configuration information in different places:
  • the hosting application’s configuration file
  • a file specified in an environment variable
  • the quartz.config file
  • the embedded configuration file

Configuring From the Hosting Application’s Configuration File

First, the factory will try to load the scheduler configuration from the <quartz> section of your application’s config file. If you are running Quartz.Net as a windows service, then the service’s configuration file will be used. This file (Quartz.Server.exe.config) can be found in the same folder as the server executable. If you are hosting the Quartz.Net scheduler in your web application, then the web.config file will be checked.

Configuring From A File Specified in an Environment Variable

If the factory was not able to load the quartz configuration section, then the second place it will check is the environment variables of the process. Specifically, the factory will check to see if the quartz.config environment variable has been defined. If such a variable exists and is not empty, the factory will try to load the configuration from whatever file is specified as the value of the variable. For example, let’s say that your configuration file is called myconfig.config. If you set the quartz.config environment variable equal to myconfig.config, then the scheduler will load the configuration information from the file named myconfig.config.

Configuring From the Quartz.config File

Let’s assume that so far, the factory has not been able to load the scheduler’s configuration. The next step that the factory will take is to try to load a file called quartz.config from the same directory where the hosting application’s assembly was loaded. In fact, this is the file that an out-of-the-box Quartz.Net server/service installation uses to configure itself.

Configuring From the Embedded Configuration File

If all of the previous configuration options fail, then the factory falls back on loading the configuration file that is embedded in the quartz assembly. In case you’re curious, here are the configuration properties that are embedded in the assembly:
Property Value
quartz.scheduler.instanceName DefaultQuartzScheduler
quartz.threadPool.type Quartz.Simpl.SimpleThreadPool, Quartz
quartz.threadPool.threadCount 10
quartz.threadPool.threadPriority Normal
quartz.jobStore.type Quartz.Simpl.RAMJobStore, Quartz
quartz.jobStore.misfireThreshold 60000
If after checking all of the above locations for configuration information, the factory was not able to configure the scheduler, then an exception is thrown. As you can probably gather from the previous paragraph, unless you actually change the quartz assembly, this configuration exception will not be raised, because a configuration file is already embedded in the quartz assembly by default.

One Final Step

You’d think that by now, the whole configuration process is finished and the scheduler has been configured successfully. Well, the default scheduler factory takes one last step before giving you the configured scheduler.
Here’s what happens just after your configuration is loaded: if any of the configuration properties that you set in a configuration are also present in the environment variables, then the factory will overwrite them with the environment value. This the expected behavior. However, due to a bug, this does not work in Quartz.Net 1.0, but is fixed in Quartz.Net 2.0.

Tuesday, February 8, 2011

New in Quartz.Net 2.0-No More IStatefulJob

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

Welcome to the the third post of my New in Quartz.Net 2.0 series.

Goodbye IStatefulJob

Yes, it’s true. In Quartz.Net 2.0 the IStatefulJob has been marked as obsolete. To be sure, it’s still available to use, but if you are using IStatefulJobs you should think about migrating them to use the new Quartz.Net 2.0 features.

The Disallow Concurrent Execution Attribute

So how do I implement an IStatefulJob in Quartz.Net 2.0? Fortunately, it’s not very hard to do. You just need to mark your IJob with the new DisallowConcurrentExecutionAttribute. Any jobs that have this attribute set are only allowed to run one instance of the job at a time. Keep in mind that an instance is defined by a JobKey (job name and group combination), which means that if you have the same job type scheduled with a different job name or a different job group, you may get more than one instance of that type of job running at the same time.

The Persist Job Data After Execution Attribute

Seems simple enough so far? Well, it is! There is only one other thing to consider. If you want the scheduler to persist the job’s state between executions, you must also add the PersistJobDataAfterExecutionAttribute to the class. This attribute tells the scheduler that you would like for it store the job’s JobDataMap once it is finished executing.

One Last Thing

While you can use each of these attributes separately, if you are going to use the PersistJobDataAfterExecutionAttribute to persist your job’s state, you should always use the DisallowConcurrentExecutionAttribute, because if you don’t, you might find that your job state is corrupted due to race conditions.

Friday, February 4, 2011

New in Quartz.Net 2.0–Directory Scan Job

NOTE: I'm now blogging at http://jayvilalta.com/blog and not updating this blog anymore. For information on the latest version of Quartz.Net, visit me there.

This is the second post of the New in Quartz.Net 2.0 series. Today we will look at the DirectoryScanJob, which is a new job in Quartz.Net 2.0.
The DirectoryScanJob is very similar to the FileScanJob, which is available in Quartz.Net 1.0. As you can probably tell from the names, the main difference between the two jobs is that the DirectoryScanJob scans for changes in… you quessed it… a directory , whereas the FileScanJob scans for changes in… yes, you guessed it again… a file. If you’re not familiar with the FileScanJob, don’t worry, we will look at the DirectoryScanJob in detail.

Introducing the DirectoryScanJob

First, let’s take a look at what the documentation says about this new job:
“Inspects a directory and compares whether any files' "last modified dates" have changed since the last time it was inspected.  If one or more files have been updated (or created), the job invokes a "call-back" method on an identified DirectoryScanListener that can be found in the SchedulerContext.”
That seems pretty self explanatory, no? Basically, you tell the job that it needs to check a certain directory to see if new files have been added or if existing files have been modified. When the job is executed it checks the files and stores the file information in its JobDataMap. The next time the job runs, it compares the information stored in the job’s JobDataMap, with the files in the directory. If any changes are found, the job calls the FilesUpdatedOrAdded method of the IDirectoryScanListener that was configured for that job, passing in an arrary of FileInfo objects.

Configuring the DirectoryScanJob

To configure and use a DirectoryScanJob, you will need to set the following keys in the JobDataMap:
  1. DIRECTORY_NAME: this is the path to the directory that you want to monitor.
  2. DIRECTORY_SCAN_LISTENER_NAME: this is the name of the DirectoryScanListener that should called if changes are detected
  3. MINIMUM_UPDATE_AGE: this is an optional parameter which lets you specify how many milliseconds must have passed since the file’s last modified time in order to consider the file modified. If this parameter is not specified, it defaults to 5000 milliseconds.

Possible Uses for the DirectoryScanJob

There are many uses for the DirectoryScanJob, so let’s list a few here:
  • You could replace any FileWatcher type services that you might have implemented
  • Scan files uploaded by users for viruses
  • Schedule other jobs to process files that have been copied to the directory for processing
The next post in the series will cover a change to the IStatefulJob interface: it has been marked as obsolete and is replaced by a class attribute!