I know that I can deny onStop feauture by setting CanStop option to false in Service's properties. This is not what I want cause this will permanently deny onStop capabilities.
What I want is to grant/deny stop capabilities programmatically. My service lifecycle is pretty simple:
starts => { run some action => sleeps for 2 minutes } x nTimes => stop
What I would is to deny stop capabilities when service is in action and grant that feature when service is idle (i.e. If user try to stop it when not permitted nothing happen, else the service really stop itself).
This is how my service is written, I have various way to understand if is idle.
partial class Service : ServiceBase
{
private static readonly log4net.ILog log = log4net.LogManager.GetLogger(System.Reflection.MethodBase.GetCurrentMethod().DeclaringType);
private DateTime _lastRun = DateTime.Now;
private System.Timers.Timer _timer = new System.Timers.Timer(Convert.ToInt32(ConfigurationManager.AppSettings["pollingInterval"]) * 60 * 1000);
public Service()
{
InitializeComponent();
}
private void timer_Elapsed(object sender, System.Timers.ElapsedEventArgs e)
{
_timer.Enabled = false;
DateTime now = DateTime.Now;
if (now.Minute > _lastRun.Minute)
{
ClientVS cvs = new ClientVS();
cvs.run();
}
_lastRun = now;
_timer.Enabled = true;
}
protected override void OnStart(string[] args)
{
log.Info("Started");
_timer.Elapsed += timer_Elapsed;
_timer.Start();
}
protected override void OnStop()
{
log.Info("Stopped");
}
}
Don't do this. Service must stop promptly (5 seconds or less) when requested to stop. If you delay or deny stop requests from SCM, you are guaranteed to annoy your users when they reboot computer. You can potentially change service configuration dynamically, but it is probably bad idea, because changing serice config frequently does not sound right.
Semantics of the service STOP command is just this -- service has to stop at the earliest safe place. Your action is different than that. What I would recommend is to implement a custom mechanism that would represent your semantics exactly (that is, do NOT reuse STOP command). There are couple of way to do this:
Any sort of IPC mechanism. A named event would probably be easiest to implement. A service will create a named event and wait for it in one of the background threads. When event is set (by your user mode application), it would do necessary checks and would call ServiceBase.Stop conditionally.
Invent a custom service control command. This is easy to do, just override ServiceBase.OnCustomCommand, make up an integer between 128 and 255, that would represent your command, and implement OnCustomCommand to call ServiceBase.Stop conditionally. To send custom command to service you can use API or sc control command. Assuming you made up 234 as you new command:
sc control MyService 234
Related
I've got a Timer that's doing a 60 second countdown. When the ticks hit 60 seconds, it stops and disposes - no problem (I think). This is run in the context of a WebApi service. I need to be able to cancel the countdown from a UI, so I've exposed a method to handle this. Since the controller is transient (thanks Luaan) and, as Daniel points out, the app pool is not predictable, I need a way to send a "cancellable" countdown to clients. Ideas anyone?
[HttpGet]
public IHttpActionResult CancelCountdown()
{
// DOES NOTHING BECAUSE THERE'S A NEW INSTANCE OF THE CONTROLLER
timer.Stop();
timer.Dispose();
return Ok();
}
private void StartCountdown()
{
// MAY BE A BAD SOLUTION BECAUSE THE APP POOL MAY RECYCLE
timer.Interval = _timeIntervalInMilliseconds;
timer.Elapsed += BroadcastToClients;
timer.Start();
}
private void BroadcastToClients(object sender, EventArgs e)
{
_elapsed += 1;
if (_elapsed == _duration)//_duration is 60
{
timer.Stop();
timer.Dispose();
return;
}
_messageHub.Clients.All.shutdown(_elapsed);
}
It's kind of hard to provide an adequate solution without knowing what you're trying to accomplish with this, but i'll give it a shot.
As Luaan pointed out, controllers are designed to be essentially stateless, so you shouldn't put instance variable on them except for it's external dependencies, since each request creates a new instance of the controller class.
You could store the timer on a static dictionary, indexed by a GUID, and return the GUID on your controller and use it as the cancellation token.
Something like:
private static Dictionary<string,Timer> timers = new Dictionary<Guid,Timer>();
public Guid StartCountdown()
{
// MAY BE A BAD SOLUTION BECAUSE THE APP POOL MAY RECYCLE
timer.Interval = _timeIntervalInMilliseconds;
timer.Elapsed += BroadcastToClients;
var guid = Guid.NewGuid().ToString();
timers.Add(guid,timer);
timer.Start();
return guid;
}
public IHttpActionResult CancelCountdown(Guid cancelationToken)
{
//If the timer no longer exist or the user supplied a wrong token
if(!timers.HasKey(cancelationToken)) return;
var timer = timers[cancelationToken];
timer.Stop();
timer.Dispose();
timers.Remove(cancelationToken);
}
However this won't solve the problem with the AppPool recycling. For a more robust solution, instead of using a timer, you could store the start date and time of each countdown in a more permanent storage (say an SQL database, a NoSQL databse, a redis server or whatever), and have a running thread or global timer, or something like Hangfire, initialized on startup, that constantly checks your countdown storage. If enough time has passed to send a broadcast message you send it, and mark the countdown as finished. If a user wants to cancel the countdown, the controller will simply read the appropiate record, mark it as cancelled, and your running thread can ignore it.
If you go with this approach, you'll need to take into account some considerations:
If the timer interval is set too short you could have a perfomance bottleneck for having to access a permament storage too often. If the interval is too long, the countdown won't be too precise.
To alleviate this problem you could store the countdowns start time in permanent storage, in case the app pool resets and you need to restore them. And also have them stored in memory on a static variable for quicker access.
Please note that if you're working with a server farm instead of a single server, static variables won't be shared across instances.
I want to create a windows service that performs some really long and heavy work. The code is inside OnStart method like this:
protected override void OnStart(string[] args)
{
System.IO.File.WriteAllText(
#"C:\MMS\Logs\WinServiceLogs.txt",
DateTime.Now + "\t MMS Service started."
);
this.RequestAdditionalTime(5*60*1000);
this.RunService();
}
this.RunService() sends a request to WCF service library hosted on IIS. It does some really long processes, ranging from 1-20 min, depending on the data it has to process. This service that I'm writing is supposed to be scheduled to run every day in the morning. So far, it runs and works fine, but when the time goes over a few seconds or min, it generates timeout exception. This causes the windows service to be in unstable state, and I can't stop or uninstall it without restarting the computer. Since, I'm trying to create an automated system, this is an issue.
I did do this.RequestAdditionalTime(), but I'm not sure whether it's doing what it's supposed to or not. I don't get the timeout error message, but now I don't know how to schedule it so it runs every day. If the exception occurs, then it won't run the next time. There were several articles and SO's I found, but there's something I'm missing and I can't understand it.
Should I create a thread? Some articles say I shouldn't put heavy programs in OnStart, where should I put the heavy codes then? Right now, when the service starts, it does this huge data processing which makes the Windows Service status to "Starting", and it stays there for long time until either the program crashes due to timeout, or completes successfully. How can I start the service, then set the status to Running while the code is running to do some data processing?
Your situation might be better suited for a scheduled task as Lloyd said in the comments above. But if you really want to use a Windows service, this is what you would need to add/update in your service code. This will allow your service to list as started and not timeout on you. You can adjust the timer length to suit your needs.
private Timer processingTimer;
public YourService()
{
InitializeComponent();
//Initialize timer
processingTimer = new Timer(60000); //Set to run every 60 seconds
processingTimer.Elapsed += processingTimer_Elapsed;
processingTimer.AutoReset = true;
processingTimer.Enabled = true;
}
private void processingTimer_Elapsed(object sender, ElapsedEventArgs e)
{
//Check the time
if (timeCheck && haventRunToday)
//Run your code
//You should probably still run this as a separate thread
this.RunService();
}
protected override void OnStart(string[] args)
{
//Start the timer
processingTimer.Start();
}
protected override void OnStop()
{
//Check to make sure that your code isn't still running... (if separate thread)
//Stop the timer
processingTimer.Stop();
}
protected override void OnPause()
{
//Stop the timer
processingTimer.Stop();
}
protected override void OnContinue()
{
//Start the timer
processingTimer.Start();
}
I'm fairly new to C#, and recently built a small webapp using .NET 4.0. This app has 2 parts: one is designed to run permanently and will continuously fetch data from given resources on the web. The other one accesses that data upon request to analyze it. I'm struggling with the first part.
My initial approach was to set up a Timer object that would execute a fetch operation (whatever that operation is doesn't really matter here) every, say, 5 minutes. I would define that timer on Application_Start and let it live after that.
However, I recently realized that applications are created / destroyed based on user requests (from my observation they seem to be destroyed after some time of inactivity). As a consequence, my background activity will stop / resume out of my control where I would like it to run continuously, with absolutely no interruption.
So here comes my question: is that achievable in a webapp? Or do I absolutely need a separate Windows service for that kind of things?
Thanks in advance for your precious help!
Guillaume
While doing this on a web app is not ideal..it is achievable, given that the site is always up.
Here's a sample: I'm creating a Cache item in the global.asax with an expiration. When it expires, an event is fired. You can fetch your data or whatever in the OnRemove() event.
Then you can set a call to a page(preferably a very small one) that will trigger code in the Application_BeginRequest that will add back the Cache item with an expiration.
global.asax:
private const string VendorNotificationCacheKey = "VendorNotification";
private const int IntervalInMinutes = 60; //Expires after X minutes & runs tasks
protected void Application_Start(object sender, EventArgs e)
{
//Set value in cache with expiration time
CacheItemRemovedCallback callback = OnRemove;
Context.Cache.Add(VendorNotificationCacheKey, DateTime.Now, null, DateTime.Now.AddMinutes(IntervalInMinutes), TimeSpan.Zero,
CacheItemPriority.Normal, callback);
}
private void OnRemove(string key, object value, CacheItemRemovedReason reason)
{
SendVendorNotification();
//Need Access to HTTPContext so cache can be re-added, so let's call a page. Application_BeginRequest will re-add the cache.
var siteUrl = ConfigurationManager.AppSettings.Get("SiteUrl");
var client = new WebClient();
client.DownloadData(siteUrl + "default.aspx");
client.Dispose();
}
private void SendVendorNotification()
{
//Do Tasks here
}
protected void Application_BeginRequest(object sender, EventArgs e)
{
//Re-add if it doesn't exist
if (HttpContext.Current.Request.Url.ToString().ToLower().Contains("default.aspx") &&
HttpContext.Current.Cache[VendorNotificationCacheKey] == null)
{
//ReAdd
CacheItemRemovedCallback callback = OnRemove;
Context.Cache.Add(VendorNotificationCacheKey, DateTime.Now, null, DateTime.Now.AddMinutes(IntervalInMinutes), TimeSpan.Zero,
CacheItemPriority.Normal, callback);
}
}
This works well, if your scheduled task is quick.
If it's a long running process..you definitely need to keep it out of your web app.
As long as the 1st request has started the application...this will keep firing every 60 minutes even if it has no visitors on the site.
I suggest putting it in a windows service. You avoid all the hoops mentioned above, the big one being IIS restarts. A windows service also has the following benefits:
Can automatically start when the server starts. If you are running in IIS and your server reboots, you have to wait until a request is made to start your process.
Can place this data fetching process on another machine if needed
If you end up load-balancing your website on multiple servers, you could accidentally have multiple data fetching processes causing you problems
Easier to main the code separately (single responsibility principle). Easier to maintain the code if it's just doing what it needs to do and not also trying to fool IIS.
Create a static class with a constructor, creating a timer event.
However like Steve Sloka mentioned, IIS has a timeout that you will have to manipulate to keep the site going.
using System.Runtime.Remoting.Messaging;
public static class Variables
{
static Variables()
{
m_wClass = new WorkerClass();
// creates and registers an event timer
m_flushTimer = new System.Timers.Timer(1000);
m_flushTimer.Elapsed += new System.Timers.ElapsedEventHandler(OnFlushTimer);
m_flushTimer.Start();
}
private static void OnFlushTimer(object o, System.Timers.ElapsedEventArgs args)
{
// determine the frequency of your update
if (System.DateTime.Now - m_timer1LastUpdateTime > new System.TimeSpan(0,1,0))
{
// call your class to do the update
m_wClass.DoMyThing();
m_timer1LastUpdateTime = System.DateTime.Now;
}
}
private static readonly System.Timers.Timer m_flushTimer;
private static System.DateTime m_timer1LastUpdateTime = System.DateTime.MinValue;
private static readonly WorkerClass m_wClass;
}
public class WorkerClass
{
public delegate WorkerClass MyDelegate();
public void DoMyThing()
{
m_test = "Hi";
m_test2 = "Bye";
//create async call to do the work
MyDelegate myDel = new MyDelegate(Execute);
AsyncCallback cb = new AsyncCallback(CommandCallBack);
IAsyncResult ar = myDel.BeginInvoke(cb, null);
}
private WorkerClass Execute()
{
//do my stuff in an async call
m_test2 = "Later";
return this;
}
public void CommandCallBack(IAsyncResult ar)
{
// this is called when your task is complete
AsyncResult asyncResult = (AsyncResult)ar;
MyDelegate myDel = (MyDelegate)asyncResult.AsyncDelegate;
WorkerClass command = myDel.EndInvoke(ar);
// command is a reference to the original class that envoked the async call
// m_test will equal "Hi"
// m_test2 will equal "Later";
}
private string m_test;
private string m_test2;
}
I think you can can achieve it by using a BackgroundWorker, but i would rather suggest you to go for a service.
Your application context lives as long as your Worker Process in IIS is functioning. In IIS there's some default timeouts for when the worker process will recycle (e.g. Number of Idle mins (20), or regular intervals (1740).
That said, if you adjust those settings in IIS, you should be able to have the requests live, however, the other answers of using a Service would work as well, just a matter of how you want to implement.
I recently made a file upload functionality for uploading Access files to the database (not the best way but just a temporary fix to a longterm issue).
I solved it by creating a background thread that ran through the ProcessAccess function, and was deleted when completed.
Unless IIS has a setting in which it kills a thread after a set amount of time regardless of inactivity, you should be able to create a thread that calls a function that never ends. Don't use recursion because the amount of open functions will eventually blow up in you face, but just have a for(;;) loop 5,000,000 times so it'll keep busy :)
Application Initialization Module for IIS 7.5 does precisely this type of init work. More details on the module are available here Application Initialization Module
At the moment im doing a chat web app where multiple users can chat and it can control multiple rooms. Its working and getting the job done.
Right now its using ajax(jquery is used) and just use GET to the server.aspx with different query parameters and then return some content.(It is meant to be build into a larger project later on)
But I have one thing that I cannot figure out how to build for it and hopin' someone had a splendid idea :)
A "Keep Alive" (or TimeToLive) service on the users. The service should ensure when a user disconnects(machine crash - Browser/window close) the user times out from the chat room.
My idea was that on every request from the user TO the server it should update a TTL list(a list with a userid and a "timestamp") and this part is easy.
Now comes my challenge
Then there should be some service running on the server that continuesly checks this TTL list to see if any stamps has run out and if it has remove the user from the room
But how and where can I do this server service in .net ? Or do you have another approch ? :)
I would just have a table called something like "LastPing" with user id and a date.
Put a piece of javascript which calls a page on your site at regular intervals (window.setInterval(...)) - that page just updates the table with the current datetime or does an insert if no rows are updated.
Finally, create a sql server job/task that selects user id from Lastping where date is older than currentdate - 30 mins (or whatever). Those user ids get deleted from any chat rooms etc. and finally removed from the LastPing table.
I think that's about it :)
You could run a Console Application (or run it as a Windows Service) that could scan your TTL list using a Timer that ticks on a set interval to process them as you wish. Could all be done in .net, preventing you from having to store your business logic in an SSIS package within SQL server.
If you're going down this path I would recommend writing a windows service that can also be run as a console app. Query the Environment.UserInteractive property to work out which version is being run - this will help with your development because a console application can be a little more verbose than a windows service.
Here is a code sample:
public partial class Service1 : ServiceBase
{
//Need to keep a reference to this object, else the Garbage Collector will clean it up and prevent further events from firing.
private System.Threading.Timer _timer;
static void Main(string[] args)
{
if (Environment.UserInteractive)
{
var service = new Service1();
Log.Debug("Starting Console Application");
service.OnStart(args);
// The service can now be accessed.
Console.WriteLine("Service ready.");
Console.WriteLine("Press <ENTER> to terminate the application.");
Console.ReadLine();
service.OnStop();
return;
}
var servicesToRun = new ServiceBase[]
{
new Service1()
};
Run(servicesToRun);
}
public Service1()
{
InitializeComponent();
}
protected override void OnStart(string[] args)
{
// For a single instance, this is a bit heavy handed, but if you're creating of a number of them
// the NT service will need to return in a short period of time and thus I use QueueUserWorkItem
ThreadPool.QueueUserWorkItem(SetupTimer, args);
}
protected override void OnStop()
{
}
private void SetupTimer(object obj)
{
//Set the emailInterval to be 1 minute by default
const int interval = 1;
//Initialize the timer, wait 5 seconds before firing, and then fire every 15 minutes
_timer = new Timer(TimerDelegate, 5000, 1, 1000 * 60 * interval);
}
private static void TimerDelegate(object stateInfo)
{
//Perform your DB TTL Check here
}
}
For that type of solution, you would need to setup a separate Thread that is periodically checking for users to expire, or utilize a library for scheduled tasks and similarly setup a scheduled task.
Needed:
A Windows Service That Executes Jobs from a Job Queue in a DB
Wanted:
Example Code, Guidance, or Best Practices for this type of Application
Background:
A user will click on an ashx link that will insert a row into the DB.
I need my windows service to periodically poll for rows in this table, and it should execute a unit of work for each row.
Emphasis:
This isn't completely new terrain for me.
EDIT: You can assume that I know how to create a Windows Service and basic data access.
But I need to write this service from scratch.
And I'd just like to know upfront what I need to consider.
EDIT: I'm most worried about jobs that fail, contention for jobs, and keeping the service running.
Given that you are dealing with a database queue, you have a fair cut of the job already done for you due to the transactional nature of databases. Typical queue driven application has a loop that does:
while(1) {
Start transction;
Dequeue item from queue;
process item;
save new state of item;
commit;
}
If processing crashes midway, the transaction rolls back and the item is processed on the next service start up.
But writing queues in a database is actually a lot trickier than you believe. If you deploy a naive approach, you'll find out that your enqueue and dequeue are blocking each other and the ashx page becomes unresponsive. Next you'll discover the dequeue vs. dequeue are deadlocking and your loop is constantly hitting error 1205. I strongly urge you to read this article Using Tables as Queues.
Your next challenge is going to be getting the pooling rate 'just right'. Too aggressive and your database will be burning hot from the pooling requests. Too lax and your queue will grow at rush hours and will drain too slowly. You should consider using an entirely different approach: use a SQL Server built-in QUEUE object and rely on the magic of the WAITFOR(RECEIVE) semantics. This allows for completely poll free self load tuning service behavior. Actually, there is more: you don't need a service to start with. See Asynchronous Procedures Execution for an explanation on what I'm talking about: launching processing asynchronously in SQL Server from a web service call, in a completely reliable manner. And finally, if the logic must be in C# process then you can leverage the External Activator, which allows the processing to be hosted in standalone processes as opposed to T-SQL procedures.
First you'll need to consider
How often to poll for
Does your service just stop and start or does it support pause and continue.
Concurrency. Services can increase the likelihood of a encountering a problem
Implementation
Use a System.Timers.Timer not a Threading.Timer
Maker sure you set the Timer.AutoReset to false. This will stop the reentrant problem.
Make sure to include execution time
Here's the basic framework of all those ideas. It includes a way to debug this which is a pain
public partial class Service : ServiceBase{
System.Timers.Timer timer;
public Service()
{
timer = new System.Timers.Timer();
//When autoreset is True there are reentrancy problme
timer.AutoReset = false;
timer.Elapsed += new System.Timers.ElapsedEventHandler(DoStuff);
}
private void DoStuff(object sender, System.Timers.ElapsedEventArgs e)
{
Collection stuff = GetData();
LastChecked = DateTime.Now;
foreach (Object item in stuff)
{
try
{
item.Dosomthing()
}
catch (System.Exception ex)
{
this.EventLog.Source = "SomeService";
this.EventLog.WriteEntry(ex.ToString());
this.Stop();
}
TimeSpan ts = DateTime.Now.Subtract(LastChecked);
TimeSpan MaxWaitTime = TimeSpan.FromMinutes(5);
if (MaxWaitTime.Subtract(ts).CompareTo(TimeSpan.Zero) > -1)
timer.Interval = MaxWaitTime.Subtract(ts).TotalMilliseconds;
else
timer.Interval = 1;
timer.Start();
}
protected override void OnPause()
{
base.OnPause();
this.timer.Stop();
}
protected override void OnContinue()
{
base.OnContinue();
this.timer.Interval = 1;
this.timer.Start();
}
protected override void OnStop()
{
base.OnStop();
this.timer.Stop();
}
protected override void OnStart(string[] args)
{
foreach (string arg in args)
{
if (arg == "DEBUG_SERVICE")
DebugMode();
}
#if DEBUG
DebugMode();
#endif
timer.Interval = 1;
timer.Start();
}
private static void DebugMode()
{
Debugger.Break();
}
}
EDIT Fixed loop in Start()
EDIT Turns out Milliseconds is not the same as TotalMilliseconds
You may want to have a look at Quartz.Net to manage scheduling the jobs. Not sure if it will fit your particular situation, but it's worth a look.
Some things I can think of, based on your edit:
Re: job failure:
Determine whether a job can be retried and do one of the following:
Move the row to an "error" table for logging / reporting later OR
Leave the row in the queue so that it will be reprocessed by the job service
You could add a column like WaitUntil or something similar to delay retrying the job after a failure
Re: contention:
Add a timestamp column such as "JobStarted" or "Locked" to track when the job was started. This will prevent other threads (assuming your service is multithreaded) from trying to execute the job simultaneously.
You'll need to have some cleanup process that goes through and clears stale jobs for re-processing (in the event the job service fails and your lock is never released).
Re: keeping the service running
You can tell windows to restart a service if it fails.
You can detect previous failure upon startup by keeping some kind of file open while the service is running and deleting it upon successful shutdown. If your service starts up and that file already exists, you know the service previously failed and can alert an operator or perform the necessary cleanup operations.
I'm really just poking around in the dark here. I'd strongly suggest prototyping the service and returning with any specific questions about the way it functions.