ASP.NET全局后台工作线程

Question

I would like to create a background thread to pull data from another server periodically and send out real time email alerts to users.

Requirement:

1. The worker should never time out.

2. The worker should be running on a separate thread which means the server should still be able to handle user's requests while it is running

3. The worker should not be created per request but rather as a global thread which will be running when the server starts

I know it is a bad idea to create a background thread per request, but what is the best practice for a global backgroundworker thread? Thanks in advance

Edited:

I am using Windows Azure to host the site, so I dont think I can create a windows service to run the task

Answer 1

Usually, you don't run such tasks in the web application itself as the application pool will be shut down after some time of inactivity depending on the configuration of the environment.

To make this work in a reliable way, create a separate application that periodically retrieves the data and sends the alerts. There are several ways to achieve this:

1. A very lightweight approach would be to create a console application and have a scheduler (eg Windows task scheduler) run it periodically.
2. A more sophisticated way is to create a Windows service that is started when the system starts and periodically executes the task.
3. If your application is integrated into a specific environment, there might already be a scheduling component available, eg in SharePoint you can implement jobs and let the Timer service run these.

Answer 2

Most people would recommend you to work with a Windows Service to accomplish it. However, a reasonable way to do this would be using a scheduling framework like Quartz .NET : http://quartznet.sourceforge.net/

That way, if you think about it, your application deployment would even be easier - no win services or EXEs to deploy.

So If you decide to do it and run it embedded in your ASP.NET application, then you can start utilize it in the Global.Asax file, at Application_Start() , like this:

NameValueCollection properties = new NameValueCollection();
properties["quartz.scheduler.instanceName"] = "RemoteServer";

// set thread pool info
properties["quartz.threadPool.type"] = "Quartz.Simpl.SimpleThreadPool, Quartz";
properties["quartz.threadPool.threadCount"] = "5";
properties["quartz.threadPool.threadPriority"] = "Normal";

ISchedulerFactory sf = new StdSchedulerFactory(properties);
IScheduler sched = sf.GetScheduler();
sched.Start();

Answer 3

I needed something similar to build scrapers. What I did is use the .Net ThreadPool class to send async http requests. I built a wrapper for building async requests with state object and then call:

ThreadPool.QueueUserWorkItem(new WaitCallback(asyncWrapper.BeginGetMethod), asyncStateObject);

Answer 4

Typically web servers are designed for serving up web pages. Requests come in, HTML pages (among other things) go out. Unless you have full control over the server, what you are trying to do will be hard to achieve.