A critical vulnerability was discovered in React Server Components (Next.js). Our systems remain protected but we advise to update packages to newest version. Learn More

Anders Hattestad
Oct 12, 2011
  4201
(4 votes)

A quick fix when saving to DSS takes a bit too much time

In a project I’m currently working on I need to update some DSS tables when a user is viewing a page. This is not a problem, but sometimes the save method takes a bit to much time.

Since the saving part of the data  not necessarily need to be done in that user thread, I have created a Lazy saver class that do the saving in a different thread.

All I needed to do is to implement my DDS classes with an interface ISaveMe and then save the item like this

Code Snippet
  1. public void SaveMe()
  2. {
  3.     Store.Save(this);
  4. }
  5. public void Save()
  6. {
  7.     LazyDSSSave.Current.AddToSave(this);
  8. }
  9. public static void Save(T item)
  10. {
  11.     LazyDSSSave.Current.AddToSave(item);
  12. }
  13. public static DynamicDataStore Store
  14. {
  15.     get
  16.     {
  17.         return DynamicDataStoreFactory.Instance.GetStore(typeof(T));
  18.     }
  19. }
Then the code bellow will start a new thread if needed an do the saving there.

It is important to remember that all exceptions in a thread that is not suppressed will make the app pool to restart. Therefore is very important to catch all errors.

Code Snippet
  1. using System;
  2. using System.Collections.Generic;
  3. using System.Linq;
  4. using System.Text;
  5. using System.Collections;
  6. using System.Threading;
  7.  
  8. namespace Itera.Data
  9. {
  10.     public interface ISaveMe
  11.     {
  12.         void SaveMe();
  13.     }
  14.     public class LazyDSSSave
  15.     {
  16.         protected Queue ActiveQueue = new Queue();
  17.         ThreadStart backgroundJob;
  18.         Thread thread;
  19.         #region Current
  20.         private static LazyDSSSave _current = new LazyDSSSave();
  21.         public static LazyDSSSave Current
  22.         {
  23.             get
  24.             {
  25.                 if (_current == null)
  26.                     _current = new LazyDSSSave();
  27.                 return _current;
  28.             }
  29.         }
  30.         #endregion
  31.  
  32.         public void AddToSave(object action)
  33.         {
  34.  
  35.             lock (ActiveQueue.SyncRoot)
  36.             {
  37.                 if (!ActiveQueue.Contains(action))
  38.                 {
  39.                     ActiveQueue.Enqueue(action);
  40.                 }
  41.             }
  42.             EnsureBackgroundJob();
  43.         }
  44.  
  45.         object lockObject = new object();
  46.         #region EnsureBackgroundJob
  47.         public void EnsureBackgroundJob()
  48.         {
  49.             lock (lockObject)
  50.             {
  51.                 if (thread == null || backgroundJob == null || !thread.IsAlive)
  52.                 {
  53.                     if (this.backgroundJob == null)
  54.                     {
  55.                         this.backgroundJob = new ThreadStart(this.Worker);
  56.                         
  57.                     }
  58.                     if (this.thread == null || !thread.IsAlive)
  59.                     {
  60.                         this.thread = new Thread(this.backgroundJob);
  61.                         this.thread.IsBackground = true;
  62.                         this.thread.Start();
  63.                     }
  64.                 }
  65.             }
  66.         }
  67.         #endregion
  68.       
  69.         #region Worker
  70.  
  71.         private void Worker()
  72.         {
  73.             try
  74.             {
  75.                 while (true)
  76.                 {
  77.                     object item = null;
  78.                     lock (ActiveQueue.SyncRoot)
  79.                     {
  80.                         if (ActiveQueue.Count == 0)
  81.                             return;
  82.                         item = ActiveQueue.Dequeue();
  83.                         
  84.                     }
  85.                     try
  86.                     {
  87.                         if (item is ISaveMe)
  88.                             (item as ISaveMe).SaveMe();
  89.                     }
  90.                     catch (Exception exception)
  91.                     {
  92.                        
  93.                         //log.Error("", exception);
  94.                     }
  95.                 }
  96.             }
  97.             catch (System.Exception error)
  98.             {
  99.                 //log.Error("", exception);
  100.             }
  101.             finally
  102.             {
  103.                 lock (lockObject)
  104.                 {
  105.                     if (backgroundJob != null)
  106.                         backgroundJob = null;
  107.                 }
  108.             }
  109.         }
  110.         #endregion
  111.     }
  112. }

 

This saves me sometimes as much as  70-100ms on each page request.

Oct 12, 2011

Comments

Magnus Rahl
Magnus Rahl Oct 12, 2011 01:15 PM

Any special reason you decided to build your own queueing/thread solution instead of using the ThreadPool?

Anders Hattestad
Anders Hattestad Oct 12, 2011 01:21 PM

Its based on the EPiServer.LazyIndexer code (at least some time ago)
I have used that way of handling this type of problems before, but I see your point :)
Its imo easier to add some statistics and logging here that can be viewed in a report later since you have one place to store these kind of data.

Please login to comment.
Latest blogs
Jhoose Security Modules v2.6.0 — Added support for Permissions Policy and .NET 10

Version 2.6.0 adds Permissions Policy header support, updates to .NET 10, improved policy management, configurable security settings, and enhanced...

Andrew Markham | Dec 6, 2025 |

Building a 360° Customer Profile With AI: How Opal + Optimizely Unlock Predictive Personalization

Creating truly relevant customer experiences requires more than collecting data—it requires understanding it. Most organizations already have rich...

Sujit Senapati | Dec 4, 2025

Building a Lightweight Optimizely SaaS CMS Solution with 11ty

Modern web development often requires striking a difficult balance between site performance and the flexibility needed by content editors. To addre...

Minesh Shah (Netcel) | Dec 3, 2025

Creating Opal Tools Using The C# SDK

Over the last few months, my colleagues at Netcel and I have partaken in two different challenge events organised by Optimizely and centered around...

Mark Stott | Dec 3, 2025