Search Results

Search found 40602 results on 1625 pages for 'event based components'.

Page 2/1625 | < Previous Page | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >

  • Event sourcing: Write event before or after updating the model

    - by Magnus
    I'm reasoning about event sourcing and often I arrive at a chicken and egg problem. Would be grateful for some hints on how to reason around this. If I execute all I/O-bound processing async (ie writing to the event log) then how do I handle, or sometimes even detect, failures? I'm using Akka Actors so processing is sequential for each event/message. I do not have any database at this time, instead I would persist all the events in an event log and then keep an aggregated state of all the events in a model stored in memory. Queries are all against this model, you can consider it to be a cache. Example Creating a new user: Validate that the user does not exist in model Persist event to journal Update model (in memory) If step 3 breaks I still have persisted my event so I can replay it at a later date. If step 2 breaks I can handle that as well gracefully. This is fine, but since step 2 is I/O-bound I figured that I should do I/O in a separate actor to free up the first actor for queries: Updating a user while allowing queries (A0 = Front end/GUI actor, A1 = Processor Actor, A2 = IO-actor, E = event bus). (A0-E-A1) Event is published to update user 'U1'. Validate that the user 'U1' exists in model (A1-A2) Persist event to journal (separate actor) (A0-E-A1-A0) Query for user 'U1' profile (A2-A1) Event is now persisted continue to update model (A0-E-A1-A0) Query for user 'U1' profile (now returns fresh data) This is appealing since queries can be processed while I/O-is churning along at it's own pace. But now I can cause myself all kinds of problems where I could have two incompatible commands (delete and then update) be persisted to the event log and crash on me when replayed up at a later date, since I do the validation before persisting the event and then update the model. My aim is to have a simple reasoning around my model (since Actor processes messages sequentially single threaded) but not be waiting for I/O-bound updates when Querying. I get the feeling I'm modeling a database which in itself is might be a problem. If things are unclear please write a comment.

    Read the article

  • Backward compatibility with event-sourcing

    - by Tomas Jansson
    How do you stay backward compatible with event-sourcing? Let say you release a version that has one kind of event, let call it X. You know how to handle that event in all the systems that extracts the events from the event source. In a later release you make a change to event X or delete it, how do you stay backward compatible with that? To have a fully functional system you need to be able to handle the old event as the same time as you need to handle the updated version. Or if you delete that event type, then you will be stuck with code that is only there to handle legacy events which in my head can be a little bit messy in the long run.

    Read the article

  • Time based movement Vs Frame rate based movement?

    - by sil3nt
    Hello there, I'm new to Game programmming and SDL, and I have been following Lazyfoo's SDL tutorials. My question is related to time based motion and frame rate based motion, basically which is better or appropriate depending on situations?. Could you give me an example where each of these methods are used?. Another question I have is that, in lazyfoo's two Motion tutorials (FPS based and time based) The time based method showed a much smoother animation while the Frame rate based one was a little hiccupy, meaning you could clearly see the gap between the previous location of the dot and its current position when you compare the two programs. As beginner which method should I stick to?(all I want is smooth animations).

    Read the article

  • component Initialization in component-based game architectures

    - by liortal
    I'm develping a 2d game (in XNA) and i've gone slightly towards a component-based approach, where i have a main game object (container) that holds different components. When implementing the needed functionality as components, i'm now faced with an issue -- who should initialize components? Are components usually passed in initialized into an entity, or some other entity initialized them? In my current design, i have an issue where the component, when created, requires knowledge regarding an attached entity, however these 2 events may not happen at the same time (component construction, attaching to a game entity). I am looking for a standard approach or examples of implementations that work, that overcome this issue or present a clear way to resolve it

    Read the article

  • How to structure game states in an entity/component-based system

    - by Eva
    I'm making a game designed with the entity-component paradigm that uses systems to communicate between components as explained here. I've reached the point in my development that I need to add game states (such as paused, playing, level start, round start, game over, etc.), but I'm not sure how to do it with my framework. I've looked at this code example on game states which everyone seems to reference, but I don't think it fits with my framework. It seems to have each state handling its own drawing and updating. My framework has a SystemManager that handles all the updating using systems. For example, here's my RenderingSystem class: public class RenderingSystem extends GameSystem { private GameView gameView_; /** * Constructor * Creates a new RenderingSystem. * @param gameManager The game manager. Used to get the game components. */ public RenderingSystem(GameManager gameManager) { super(gameManager); } /** * Method: registerGameView * Registers gameView into the RenderingSystem. * @param gameView The game view registered. */ public void registerGameView(GameView gameView) { gameView_ = gameView; } /** * Method: triggerRender * Adds a repaint call to the event queue for the dirty rectangle. */ public void triggerRender() { Rectangle dirtyRect = new Rectangle(); for (GameObject object : getRenderableObjects()) { GraphicsComponent graphicsComponent = object.getComponent(GraphicsComponent.class); dirtyRect.add(graphicsComponent.getDirtyRect()); } gameView_.repaint(dirtyRect); } /** * Method: renderGameView * Renders the game objects onto the game view. * @param g The graphics object that draws the game objects. */ public void renderGameView(Graphics g) { for (GameObject object : getRenderableObjects()) { GraphicsComponent graphicsComponent = object.getComponent(GraphicsComponent.class); if (!graphicsComponent.isVisible()) continue; GraphicsComponent.Shape shape = graphicsComponent.getShape(); BoundsComponent boundsComponent = object.getComponent(BoundsComponent.class); Rectangle bounds = boundsComponent.getBounds(); g.setColor(graphicsComponent.getColor()); if (shape == GraphicsComponent.Shape.RECTANGULAR) { g.fill3DRect(bounds.x, bounds.y, bounds.width, bounds.height, true); } else if (shape == GraphicsComponent.Shape.CIRCULAR) { g.fillOval(bounds.x, bounds.y, bounds.width, bounds.height); } } } /** * Method: getRenderableObjects * @return The renderable game objects. */ private HashSet<GameObject> getRenderableObjects() { return gameManager.getGameObjectManager().getRelevantObjects( getClass()); } } Also all the updating in my game is event-driven. I don't have a loop like theirs that simply updates everything at the same time. I like my framework because it makes it easy to add new GameObjects, but doesn't have the problems some component-based designs encounter when communicating between components. I would hate to chuck it just to get pause to work. Is there a way I can add game states to my game without removing the entity-component design? Does the game state example actually fit my framework, and I'm just missing something? EDIT: I might not have explained my framework well enough. My components are just data. If I was coding in C++, they'd probably be structs. Here's an example of one: public class BoundsComponent implements GameComponent { /** * The position of the game object. */ private Point pos_; /** * The size of the game object. */ private Dimension size_; /** * Constructor * Creates a new BoundsComponent for a game object with initial position * initialPos and initial size initialSize. The position and size combine * to make up the bounds. * @param initialPos The initial position of the game object. * @param initialSize The initial size of the game object. */ public BoundsComponent(Point initialPos, Dimension initialSize) { pos_ = initialPos; size_ = initialSize; } /** * Method: getBounds * @return The bounds of the game object. */ public Rectangle getBounds() { return new Rectangle(pos_, size_); } /** * Method: setPos * Sets the position of the game object to newPos. * @param newPos The value to which the position of the game object is * set. */ public void setPos(Point newPos) { pos_ = newPos; } } My components do not communicate with each other. Systems handle inter-component communication. My systems also do not communicate with each other. They have separate functionality and can easily be kept separate. The MovementSystem doesn't need to know what the RenderingSystem is rendering to move the game objects correctly; it just need to set the right values on the components, so that when the RenderingSystem renders the game objects, it has accurate data. The game state could not be a system, because it needs to interact with the systems rather than the components. It's not setting data; it's determining which functions need to be called. A GameStateComponent wouldn't make sense because all the game objects share one game state. Components are what make up objects and each one is different for each different object. For example, the game objects cannot have the same bounds. They can have overlapping bounds, but if they share a BoundsComponent, they're really the same object. Hopefully, this explanation makes my framework less confusing.

    Read the article

  • Doing powerups in a component-based system

    - by deft_code
    I'm just starting really getting my head around component based design. I don't know what the "right" way to do this is. Here's the scenario. The player can equip a shield. The the shield is drawn as bubble around the player, it has a separate collision shape, and reduces the damage the player receives from area effects. How is such a shield architected in a component based game? Where I get confused is that the shield obviously has three components associated with it. Damage reduction / filtering A sprite A collider. To make it worse different shield variations could have even more behaviors, all of which could be components: boost player maximum health health regen projectile deflection etc Am I overthinking this? Should the shield just be a super component? I really think this is wrong answer. So if you think this is the way to go please explain. Should the shield be its own entity that tracks the location of the player? That might make it hard to implement the damage filtering. It also kinda blurs the lines between attached components and entities. Should the shield be a component that houses other components? I've never seen or heard of anything like this, but maybe it's common and I'm just not deep enough yet. Should the shield just be a set of components that get added to the player? Possibly with an extra component to manage the others, e.g. so they can all be removed as a group. (accidentally leave behind the damage reduction component, now that would be fun). Something else that's obvious to someone with more component experience?

    Read the article

  • Doing powerups in a component-based system

    - by deft_code
    I'm just starting really getting my head around component based design. I don't know what the "right" way to do this is. Here's the scenario. The player can equip a shield. The the shield is drawn as bubble around the player, it has a separate collision shape, and reduces the damage the player receives from area effects. How is such a shield architected in a component based game? Where I get confused is that the shield obviously has three components associated with it. Damage reduction / filtering A sprite A collider. To make it worse different shield variations could have even more behaviors, all of which could be components: boost player maximum health health regen projectile deflection etc Am I overthinking this? Should the shield just be a super component? I really think this is wrong answer. So if you think this is the way to go please explain. Should the shield be its own entity that tracks the location of the player? That might make it hard to implement the damage filtering. It also kinda blurs the lines between attached components and entities. Should the shield be a component that houses other components? I've never seen or heard of anything like this, but maybe it's common and I'm just not deep enough yet. Should the shield just be a set of components that get added to the player? Possibly with an extra component to manage the others, e.g. so they can all be removed as a group. (accidentally leave behind the damage reduction component, now that would be fun). Something else that's obvious to someone with more component experience?

    Read the article

  • Handling game logic events by behavior components

    - by chehob
    My question continues on topic discussed here I have tried implementing attribute/behavior design and here is a quick example demonstrating the issue. class HealthAttribute : public ActorAttribute { public: HealthAttribute( float val ) : mValue( val ) { } float Get( void ) const { return mValue; } void Set( float val ) { mValue = val; } private: float mValue; }; class HealthBehavior : public ActorBehavior { public: HealthBehavior( shared_ptr< HealthAttribute > health ) : pHealth( health ) { // Set OnDamage() to listen for game logic event "DamageEvent" } void OnDamage( IEventDataPtr pEventData ) { // Check DamageEvent target entity // ( compare my entity ID with event's target entity ID ) // If not my entity, do nothing // Else, modify health attribute with received DamageEvent data } protected: shared_ptr< HealthAttribute > pHealth; }; My question - is it possible to get rid of this annoying check for game logic events? In the current implementation when some entity must receive damage, game logic just fires off event that contains damage value and the entity id which should receive that damage. And all HealthBehaviors are subscribed to the DamageEvent type, which leads to any entity possesing HealthBehavior call OnDamage() even if he is not the addressee.

    Read the article

  • How to prioritize related game entity components?

    - by Paul Manta
    I want to make a game where you have to run over a bunch of zombies with your car. When moving around, the zombies have a few things to take into consideration: When there's no player around they might just roam about randomly. And even when some other component dictates a specific direction, they should wobble to the left and right randomly (like drunk people). This implies a small, random, deviation in their movement. They should avoid static obstacles. When they see they are headed towards a wall, they should reorient themselves. They should avoid the car. They should try to predict where the car will be based on its velocity and try to move out of the way. When they can, they should try to get near the player. All these types of decisions they have to do seem like they should be implemented in different components. But how should I manage them? How can I give different components different weights that reflect the importance of each decision (in a given situation)? I would need some other component that acts as a manager, but do you have any tips on how I should implement it? Or maybe there's a better solution?...

    Read the article

  • Managing game objects/components

    - by Xeon06
    Good day everyone, By far the biggest problem that has always dawned on my when programming games is how to structure my code. It just becomes an incredible mess after a while. The reason for that is because I have no idea how different classes should interact with each other. Let's have an example. Say I have a class Player, a class PlayerInput and a class Map. The player class contains information as to the location of the player, whereas the player input class handles changing that location, but by first making sure it's within a walkable area from the map class. How to structure this? My usual approach is to pass those components as parameters in the constructors of the parameters that need them, like so: var map = new Map(); var player = new Player(); var input = new PlayerInput(player, map); The problem with that is that it quickly gets messy, when you add new components you have to go through your constructors and update them, and it doesn't work well if you have mirroring references: var physics = new Physics(input); //Oops, doesn't work var input = new Input(physics); So, how do you guys usually manage this? Thanks.

    Read the article

  • Component based game engine issue

    - by Mathias Hölzl
    We are just switching from a hierarchy based game engine to a component based game engine. My problem is that when I load a model which has has a hierarchy of meshes and the way I understand is that a entity in a component based system can not have multiple components of the same type, but I need a "meshComponent" for each mesh in a model. So how could I solve this problem. On this side they implemented a Component based game engine: http://cowboyprogramming.com/2007/01/05/evolve-your-heirachy/

    Read the article

  • wxPython - Running built in event prior to custom bound event

    - by jhaukur
    I'm trying to figure out how I can specify that the mouse_down event in wxPython (StyledTextCtrl) is first handled by the built in event listener, which changes the caret position, and then handled by my own custom event handler. To bind the custom event handler I use wx.EVT_LEFT_DOWN(self.styCtrl, self.OnMouseClick) def OnMouseClick(self, evt): evt.Skip() foo() I want the built in event handler to fire and complete prior to foo().

    Read the article

  • Cutom event dispatchment location

    - by Martino Wullems
    Hello, I've been looking into custom event (listeners) for quite some time, but never succeeded in making one. There are so many different mehods, extending the Event class, but also Extending the EventDispatcher class, very confusing! I want to settle with this once and for all and learn the appriopate technique. package{ import flash.events.Event; public class CustomEvent extends Event{ public static const TEST:String = 'test'; //what exac is the purpose of the value in the string? public var data:Object; public function CustomEvent(type:String, bubbles:Boolean = false, cancelable:Boolean = false, data:Object = null):void { this.data = data; super(); } } } As far as I know a custom class where you set the requirements for the event to be dispatched has to be made: package { import flash.display.MovieClip; public class TestClass extends MovieClip { public function TestClass():void { if (ConditionForHoldToComplete == true) { dispatchEvent(new Event(CustomEvent.TEST)); } } } } I'm not sure if this is correct, but it should be something along the lines of this. Now What I want is something like a mouseevent, which can be applied to a target and does not require a specific class. It would have to work something like this: package com.op_pad._events{ import flash.events.MouseEvent; import flash.utils.Timer; import flash.events.TimerEvent; import flash.events.EventDispatcher; import flash.events.Event; public class HoldEvent extends Event { public static const HOLD_COMPLETE:String = "hold completed"; var timer:Timer; public function SpriteEvent(type:String, bubbles:Boolean=true, cancelable:Boolean=false) { super( type, bubbles, cancelable ); timer = new Timer(1000, 1); //somehow find the target where is event is placed upon -> target.addEventlistener target.addEventListener(MouseEvent.MOUSE_DOWN, startTimer); target.addEventListener(MouseEvent.MOUSE_UP, stopTimer); } public override function clone():Event { return new SpriteEvent(type, bubbles, cancelable); } public override function toString():String { return formatToString("MovieEvent", "type", "bubbles", "cancelable", "eventPhase"); } ////////////////////////////////// ///// c o n d i t i o n s ///// ////////////////////////////////// private function startTimer(e:MouseEvent):void { timer.start(); timer.addEventListener(TimerEvent.TIMER_COMPLETE, complete); } private function stopTimer(e:MouseEvent):void { timer.stop() } public function complete(e:TimerEvent):void { dispatchEvent(new HoldEvent(HoldEvent.HOLD_COMPLETE)); } } } This obviously won't work, but should give you an idea of what I want to achieve. This should be possible because mouseevent can be applied to about everything.The main problem is that I don't know where I should set the requirements for the event to be executed to be able to apply it to movieclips and sprites. Thanks in advance

    Read the article

  • Option Trading: Getting the most out of the event session options

    - by extended_events
    You can control different aspects of how an event session behaves by setting the event session options as part of the CREATE EVENT SESSION DDL. The default settings for the event session options are designed to handle most of the common event collection situations so I generally recommend that you just use the defaults. Like everything in the real world though, there are going to be a handful of “special cases” that require something different. This post focuses on identifying the special cases and the correct use of the options to accommodate those cases. There is a reason it’s called Default The default session options specify a total event buffer size of 4 MB with a 30 second latency. Translating this into human terms; this means that our default behavior is that the system will start processing events from the event buffer when we reach about 1.3 MB of events or after 30 seconds, which ever comes first. Aside: What’s up with the 1.3 MB, I thought you said the buffer was 4 MB?The Extended Events engine takes the total buffer size specified by MAX_MEMORY (4MB by default) and divides it into 3 equally sized buffers. This is done so that a session can be publishing events to one buffer while other buffers are being processed. There are always at least three buffers; how to get more than three is covered later. Using this configuration, the Extended Events engine can “keep up” with most event sessions on standard workloads. Why is this? The fact is that most events are small, really small; on the order of a couple hundred bytes. Even when you start considering events that carry dynamically sized data (eg. binary, text, etc.) or adding actions that collect additional data, the total size of the event is still likely to be pretty small. This means that each buffer can likely hold thousands of events before it has to be processed. When the event buffers are finally processed there is an economy of scale achieved since most targets support bulk processing of the events so they are processed at the buffer level rather than the individual event level. When all this is working together it’s more likely that a full buffer will be processed and put back into the ready queue before the remaining buffers (remember, there are at least three) are full. I know what you’re going to say: “My server is exceptional! My workload is so massive it defies categorization!” OK, maybe you weren’t going to say that exactly, but you were probably thinking it. The point is that there are situations that won’t be covered by the Default, but that’s a good place to start and this post assumes you’ve started there so that you have something to look at in order to determine if you do have a special case that needs different settings. So let’s get to the special cases… What event just fired?! How about now?! Now?! If you believe the commercial adage from Heinz Ketchup (Heinz Slow Good Ketchup ad on You Tube), some things are worth the wait. This is not a belief held by most DBAs, particularly DBAs who are looking for an answer to a troubleshooting question fast. If you’re one of these anxious DBAs, or maybe just a Program Manager doing a demo, then 30 seconds might be longer than you’re comfortable waiting. If you find yourself in this situation then consider changing the MAX_DISPATCH_LATENCY option for your event session. This option will force the event buffers to be processed based on your time schedule. This option only makes sense for the asynchronous targets since those are the ones where we allow events to build up in the event buffer – if you’re using one of the synchronous targets this option isn’t relevant. Avoid forgotten events by increasing your memory Have you ever had one of those days where you keep forgetting things? That can happen in Extended Events too; we call it dropped events. In order to optimizes for server performance and help ensure that the Extended Events doesn’t block the server if to drop events that can’t be published to a buffer because the buffer is full. You can determine if events are being dropped from a session by querying the dm_xe_sessions DMV and looking at the dropped_event_count field. Aside: Should you care if you’re dropping events?Maybe not – think about why you’re collecting data in the first place and whether you’re really going to miss a few dropped events. For example, if you’re collecting query duration stats over thousands of executions of a query it won’t make a huge difference to miss a couple executions. Use your best judgment. If you find that your session is dropping events it means that the event buffer is not large enough to handle the volume of events that are being published. There are two ways to address this problem. First, you could collect fewer events – examine you session to see if you are over collecting. Do you need all the actions you’ve specified? Could you apply a predicate to be more specific about when you fire the event? Assuming the session is defined correctly, the next option is to change the MAX_MEMORY option to a larger number. Picking the right event buffer size might take some trial and error, but a good place to start is with the number of dropped events compared to the number you’ve collected. Aside: There are three different behaviors for dropping events that you specify using the EVENT_RETENTION_MODE option. The default is to allow single event loss and you should stick with this setting since it is the best choice for keeping the impact on server performance low.You’ll be tempted to use the setting to not lose any events (NO_EVENT_LOSS) – resist this urge since it can result in blocking on the server. If you’re worried that you’re losing events you should be increasing your event buffer memory as described in this section. Some events are too big to fail A less common reason for dropping an event is when an event is so large that it can’t fit into the event buffer. Even though most events are going to be small, you might find a condition that occasionally generates a very large event. You can determine if your session is dropping large events by looking at the dm_xe_sessions DMV once again, this time check the largest_event_dropped_size. If this value is larger than the size of your event buffer [remember, the size of your event buffer, by default, is max_memory / 3] then you need a large event buffer. To specify a large event buffer you set the MAX_EVENT_SIZE option to a value large enough to fit the largest event dropped based on data from the DMV. When you set this option the Extended Events engine will create two buffers of this size to accommodate these large events. As an added bonus (no extra charge) the large event buffer will also be used to store normal events in the cases where the normal event buffers are all full and waiting to be processed. (Note: This is just a side-effect, not the intended use. If you’re dropping many normal events then you should increase your normal event buffer size.) Partitioning: moving your events to a sub-division Earlier I alluded to the fact that you can configure your event session to use more than the standard three event buffers – this is called partitioning and is controlled by the MEMORY_PARTITION_MODE option. The result of setting this option is fairly easy to explain, but knowing when to use it is a bit more art than science. First the science… You can configure partitioning in three ways: None, Per NUMA Node & Per CPU. This specifies the location where sets of event buffers are created with fairly obvious implication. There are rules we follow for sub-dividing the total memory (specified by MAX_MEMORY) between all the event buffers that are specific to the mode used: None: 3 buffers (fixed)Node: 3 * number_of_nodesCPU: 2.5 * number_of_cpus Here are some examples of what this means for different Node/CPU counts: Configuration None Node CPU 2 CPUs, 1 Node 3 buffers 3 buffers 5 buffers 6 CPUs, 2 Node 3 buffers 6 buffers 15 buffers 40 CPUs, 5 Nodes 3 buffers 15 buffers 100 buffers   Aside: Buffer size on multi-processor computersAs the number of Nodes or CPUs increases, the size of the event buffer gets smaller because the total memory is sub-divided into more pieces. The defaults will hold up to this for a while since each buffer set is holding events only from the Node or CPU that it is associated with, but at some point the buffers will get too small and you’ll either see events being dropped or you’ll get an error when you create your session because you’re below the minimum buffer size. Increase the MAX_MEMORY setting to an appropriate number for the configuration. The most likely reason to start partitioning is going to be related to performance. If you notice that running an event session is impacting the performance of your server beyond a reasonably expected level [Yes, there is a reasonably expected level of work required to collect events.] then partitioning might be an answer. Before you partition you might want to check a few other things: Is your event retention set to NO_EVENT_LOSS and causing blocking? (I told you not to do this.) Consider changing your event loss mode or increasing memory. Are you over collecting and causing more work than necessary? Consider adding predicates to events or removing unnecessary events and actions from your session. Are you writing the file target to the same slow disk that you use for TempDB and your other high activity databases? <kidding> <not really> It’s always worth considering the end to end picture – if you’re writing events to a file you can be impacted by I/O, network; all the usual stuff. Assuming you’ve ruled out the obvious (and not so obvious) issues, there are performance conditions that will be addressed by partitioning. For example, it’s possible to have a successful event session (eg. no dropped events) but still see a performance impact because you have many CPUs all attempting to write to the same free buffer and having to wait in line to finish their work. This is a case where partitioning would relieve the contention between the different CPUs and likely reduce the performance impact cause by the event session. There is no DMV you can check to find these conditions – sorry – that’s where the art comes in. This is  largely a matter of experimentation. On the bright side you probably won’t need to to worry about this level of detail all that often. The performance impact of Extended Events is significantly lower than what you may be used to with SQL Trace. You will likely only care about the impact if you are trying to set up a long running event session that will be part of your everyday workload – sessions used for short term troubleshooting will likely fall into the “reasonably expected impact” category. Hey buddy – I think you forgot something OK, there are two options I didn’t cover: STARTUP_STATE & TRACK_CAUSALITY. If you want your event sessions to start automatically when the server starts, set the STARTUP_STATE option to ON. (Now there is only one option I didn’t cover.) I’m going to leave causality for another post since it’s not really related to session behavior, it’s more about event analysis. - Mike Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Option Trading: Getting the most out of the event session options

    - by extended_events
    You can control different aspects of how an event session behaves by setting the event session options as part of the CREATE EVENT SESSION DDL. The default settings for the event session options are designed to handle most of the common event collection situations so I generally recommend that you just use the defaults. Like everything in the real world though, there are going to be a handful of “special cases” that require something different. This post focuses on identifying the special cases and the correct use of the options to accommodate those cases. There is a reason it’s called Default The default session options specify a total event buffer size of 4 MB with a 30 second latency. Translating this into human terms; this means that our default behavior is that the system will start processing events from the event buffer when we reach about 1.3 MB of events or after 30 seconds, which ever comes first. Aside: What’s up with the 1.3 MB, I thought you said the buffer was 4 MB?The Extended Events engine takes the total buffer size specified by MAX_MEMORY (4MB by default) and divides it into 3 equally sized buffers. This is done so that a session can be publishing events to one buffer while other buffers are being processed. There are always at least three buffers; how to get more than three is covered later. Using this configuration, the Extended Events engine can “keep up” with most event sessions on standard workloads. Why is this? The fact is that most events are small, really small; on the order of a couple hundred bytes. Even when you start considering events that carry dynamically sized data (eg. binary, text, etc.) or adding actions that collect additional data, the total size of the event is still likely to be pretty small. This means that each buffer can likely hold thousands of events before it has to be processed. When the event buffers are finally processed there is an economy of scale achieved since most targets support bulk processing of the events so they are processed at the buffer level rather than the individual event level. When all this is working together it’s more likely that a full buffer will be processed and put back into the ready queue before the remaining buffers (remember, there are at least three) are full. I know what you’re going to say: “My server is exceptional! My workload is so massive it defies categorization!” OK, maybe you weren’t going to say that exactly, but you were probably thinking it. The point is that there are situations that won’t be covered by the Default, but that’s a good place to start and this post assumes you’ve started there so that you have something to look at in order to determine if you do have a special case that needs different settings. So let’s get to the special cases… What event just fired?! How about now?! Now?! If you believe the commercial adage from Heinz Ketchup (Heinz Slow Good Ketchup ad on You Tube), some things are worth the wait. This is not a belief held by most DBAs, particularly DBAs who are looking for an answer to a troubleshooting question fast. If you’re one of these anxious DBAs, or maybe just a Program Manager doing a demo, then 30 seconds might be longer than you’re comfortable waiting. If you find yourself in this situation then consider changing the MAX_DISPATCH_LATENCY option for your event session. This option will force the event buffers to be processed based on your time schedule. This option only makes sense for the asynchronous targets since those are the ones where we allow events to build up in the event buffer – if you’re using one of the synchronous targets this option isn’t relevant. Avoid forgotten events by increasing your memory Have you ever had one of those days where you keep forgetting things? That can happen in Extended Events too; we call it dropped events. In order to optimizes for server performance and help ensure that the Extended Events doesn’t block the server if to drop events that can’t be published to a buffer because the buffer is full. You can determine if events are being dropped from a session by querying the dm_xe_sessions DMV and looking at the dropped_event_count field. Aside: Should you care if you’re dropping events?Maybe not – think about why you’re collecting data in the first place and whether you’re really going to miss a few dropped events. For example, if you’re collecting query duration stats over thousands of executions of a query it won’t make a huge difference to miss a couple executions. Use your best judgment. If you find that your session is dropping events it means that the event buffer is not large enough to handle the volume of events that are being published. There are two ways to address this problem. First, you could collect fewer events – examine you session to see if you are over collecting. Do you need all the actions you’ve specified? Could you apply a predicate to be more specific about when you fire the event? Assuming the session is defined correctly, the next option is to change the MAX_MEMORY option to a larger number. Picking the right event buffer size might take some trial and error, but a good place to start is with the number of dropped events compared to the number you’ve collected. Aside: There are three different behaviors for dropping events that you specify using the EVENT_RETENTION_MODE option. The default is to allow single event loss and you should stick with this setting since it is the best choice for keeping the impact on server performance low.You’ll be tempted to use the setting to not lose any events (NO_EVENT_LOSS) – resist this urge since it can result in blocking on the server. If you’re worried that you’re losing events you should be increasing your event buffer memory as described in this section. Some events are too big to fail A less common reason for dropping an event is when an event is so large that it can’t fit into the event buffer. Even though most events are going to be small, you might find a condition that occasionally generates a very large event. You can determine if your session is dropping large events by looking at the dm_xe_sessions DMV once again, this time check the largest_event_dropped_size. If this value is larger than the size of your event buffer [remember, the size of your event buffer, by default, is max_memory / 3] then you need a large event buffer. To specify a large event buffer you set the MAX_EVENT_SIZE option to a value large enough to fit the largest event dropped based on data from the DMV. When you set this option the Extended Events engine will create two buffers of this size to accommodate these large events. As an added bonus (no extra charge) the large event buffer will also be used to store normal events in the cases where the normal event buffers are all full and waiting to be processed. (Note: This is just a side-effect, not the intended use. If you’re dropping many normal events then you should increase your normal event buffer size.) Partitioning: moving your events to a sub-division Earlier I alluded to the fact that you can configure your event session to use more than the standard three event buffers – this is called partitioning and is controlled by the MEMORY_PARTITION_MODE option. The result of setting this option is fairly easy to explain, but knowing when to use it is a bit more art than science. First the science… You can configure partitioning in three ways: None, Per NUMA Node & Per CPU. This specifies the location where sets of event buffers are created with fairly obvious implication. There are rules we follow for sub-dividing the total memory (specified by MAX_MEMORY) between all the event buffers that are specific to the mode used: None: 3 buffers (fixed)Node: 3 * number_of_nodesCPU: 2.5 * number_of_cpus Here are some examples of what this means for different Node/CPU counts: Configuration None Node CPU 2 CPUs, 1 Node 3 buffers 3 buffers 5 buffers 6 CPUs, 2 Node 3 buffers 6 buffers 15 buffers 40 CPUs, 5 Nodes 3 buffers 15 buffers 100 buffers   Aside: Buffer size on multi-processor computersAs the number of Nodes or CPUs increases, the size of the event buffer gets smaller because the total memory is sub-divided into more pieces. The defaults will hold up to this for a while since each buffer set is holding events only from the Node or CPU that it is associated with, but at some point the buffers will get too small and you’ll either see events being dropped or you’ll get an error when you create your session because you’re below the minimum buffer size. Increase the MAX_MEMORY setting to an appropriate number for the configuration. The most likely reason to start partitioning is going to be related to performance. If you notice that running an event session is impacting the performance of your server beyond a reasonably expected level [Yes, there is a reasonably expected level of work required to collect events.] then partitioning might be an answer. Before you partition you might want to check a few other things: Is your event retention set to NO_EVENT_LOSS and causing blocking? (I told you not to do this.) Consider changing your event loss mode or increasing memory. Are you over collecting and causing more work than necessary? Consider adding predicates to events or removing unnecessary events and actions from your session. Are you writing the file target to the same slow disk that you use for TempDB and your other high activity databases? <kidding> <not really> It’s always worth considering the end to end picture – if you’re writing events to a file you can be impacted by I/O, network; all the usual stuff. Assuming you’ve ruled out the obvious (and not so obvious) issues, there are performance conditions that will be addressed by partitioning. For example, it’s possible to have a successful event session (eg. no dropped events) but still see a performance impact because you have many CPUs all attempting to write to the same free buffer and having to wait in line to finish their work. This is a case where partitioning would relieve the contention between the different CPUs and likely reduce the performance impact cause by the event session. There is no DMV you can check to find these conditions – sorry – that’s where the art comes in. This is  largely a matter of experimentation. On the bright side you probably won’t need to to worry about this level of detail all that often. The performance impact of Extended Events is significantly lower than what you may be used to with SQL Trace. You will likely only care about the impact if you are trying to set up a long running event session that will be part of your everyday workload – sessions used for short term troubleshooting will likely fall into the “reasonably expected impact” category. Hey buddy – I think you forgot something OK, there are two options I didn’t cover: STARTUP_STATE & TRACK_CAUSALITY. If you want your event sessions to start automatically when the server starts, set the STARTUP_STATE option to ON. (Now there is only one option I didn’t cover.) I’m going to leave causality for another post since it’s not really related to session behavior, it’s more about event analysis. - Mike Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Event ID: 861 - The Windows Firewall has detected an application listening for incoming traffic

    - by Chris Marisic
    Firstly, my machines aren't compromised any person suggesting such will be DV'd. The security logs on some of my networks client machines (all Windows Xp Sp3) get filled with these useless error messages. Security Failure Audit Detailed Tracking Event ID: 861 User: NT AUTHORITY\NETWORK SERVICE The Windows Firewall has detected an application listening for incoming traffic. Name: - Path: C:\WINDOWS\system32\svchost.exe Process identifier: 976 User account: NETWORK SERVICE User domain: NT AUTHORITY Service: Yes RPC server: No IP version: IPv4 IP protocol: UDP Port number: 55035 Allowed: No User notified: No It's always on various random ports of UDP so setting up a port exception isn't really an option. It's always from svchost or lsass both of which are running services from DLLs. One of the most offending processes seems to the be DnsCache. I have in my global policy under AT < Network < Network Connection < Widnows Firewall < Domain Profile (I haven't changed any standard profile options do both need configured? To allow remote administration and desktop exceptions and have a custom program exception list that has %SystemRoot%\system32\svchost.exe:*:enabled:svchost (Windows won't allow you to add this exception on a local machine but it let me have it on here in the global policy it just doesn't seem to do anything) %SystemRoot%\system32\lsass.exe:*enabled:lsass (I think this one ended all of my LSASS messages) %SystemRoot%\system32\dnsrslvr.dll:*:enabled:dnscache (I tried adding the dll itself to the exception list, this didn't seem to do anything) Is there really any other options left other than disabling the Windows Firewall entirely, disabling auditing entirely or just changing the event viewer to just auto overwrite when needed? I'd much rather fix the problem and get rid of these entries ever being created instead of just trying to cover up the problem.

    Read the article

  • how to re-enable default after doing event.preventDefault()

    - by Matt
    Hi, I know this exact question was asked here, but the answer didn't work for what I needed to do so I figured I'd give some example code and explain a bit... $(document).keypress( function (event) { // Pressing Up or Right: Advance to next video if (event.keyCode == 40 || event.keyCode == 39) { event.preventDefault(); $(".current").next().click(); } // Pressing Down or Left: Back to previous video else if (event.keyCode == 38 || event.keyCode == 37) { event.preventDefault(); $(".current").prev().click(); } } ); It basically disables the arrow keys to use them for something else, but doing: $(document).keypress(function () { }); doesn't enable the default function again... I need it to scroll the page without having to create a scroll function for it... Any ideas? Thanks, Matt

    Read the article

  • DDD North 3 Presentation and source code &ndash; &lsquo;Event Store - an introduction to a DSD for event sourcing and notifications&rsquo;

    - by Liam Westley
    Originally posted on: http://geekswithblogs.net/twickers/archive/2013/10/15/ddd-north-3-presentation-and-source-code-ndash-lsquoevent-store.aspxThank you everyone at DDD North Thanks to all the people who helped organise the cracking conference that is DDD North 3, returning to Sunderland, and the great facilities at the University of Sunderland, and the fine drinks reception at Sunderland Software City.  The whole event wouldn’t be possible without the sponsors who ensured over 400 people were kept fed and watered so they could enjoy the impressive range of sessions. And lastly, a thank you to all those delegates who gave up their free time on a Saturday to spend a day dashing between lecture rooms, including a late change to my room which saw 40 people having to brave a journey between buildings in the fine drizzle. The enthusiasm from the delegates always helps recharge my geek batteries. Presentation and source code My presentation, source code, Event Store runners and text files containing the various command line parameters used for curl is now available on GitHub; https://github.com/westleyl/DDDNorth3-EventStore. Don’t worry if you don’t have a GitHub account, you don’t need one, you can just click on the Download Zip button on the right hand menu to download all the files as a single ZIP file.  If all you want is the PowerPoint presentation, go to https://github.com/westleyl/DDDNorth3-EventStore/blob/master/Powerpoint/DDDNorth-EventStore.pptx, and click on the View Raw button. Downloading and installing Event Store and Tools Download Event Store http://download.geteventstore.com – I unzipped these files into C:\EventStore\v2.0.1 Download Curl from http://curl.haxx.se/download.html – I downloaded Win64 Generic (with SSL) and unzipped these files into C:\curl version 7.31.0 Running the tools I used in my presentation Demonstration 1 (running Event Store) You can use one of my Event Store runner command files to run the single node version of Event Store, using default ports of 2213 for HTTP and 1113  for TCP, and with a wildcard HTTP pattern.  Both take a single command line parameter to specify the location of the data and log files.  The runners assume the single node executable is located in C:\EventStore\v2.0.1, and will placed data files and logs beneath C:\EventStore\Data, i.e. RunEventStore.cmd TestData1 This will create data files in C:\EventStore\Data\TestData1\Data and log files in C:\EventStore\Data\TestData1\logs. If, when running Event Store you may see the following message, [03288,15,06:23:00.622] Failed to start http server Access is denied You will either need to run Event Store in an administrator console window, or you can use the netsh command to create a firewall permission to allow HTTP listening (this will need to be run, once, in an administrator console window), netsh http add urlacl url=http://*:2213/ user=liam You can always delete this later by running the delete; netsh http delete urlacl url=http://*:2213/ If you want to confirm that everything is running OK, open the management console in a browser by navigating to http://127.0.0.1:2213. If at any point you are asked for a user name and password use the default of ‘admin’/‘changeit’.   Demonstration 2 (reading and adding data, curl) In my second demonstration I used curl directly from the console to read streams, write events and then read back those events. On GitHub I have included is a set of curl commands, CurlCommandLine.txt, and a sample data file, SampleData.json, to load an event into a DDDNorth3 stream. As there is not much data in the Event Store at this point I used the $stats-127.0.0.1:2113 which is a stream containing performance statistics for Event Store and is updated every 30 seconds (default). Demonstration 3 (projections) On GitHub I have included a sample projection, Projection-ByRoom.txt, which will create streams based on the room on which a session was held on the DDDNorth3 agenda. Browse to the management console, http://127.0.0.1:2213.  Click on Projections, New Projection, give it a name, Sessions-ByRoom, and copy in the JavaScript in the Projection-ByRoom.txt file.  Select Continuous, tick Emit Enabled and then click on Post. It should run immediately. You may by challenged for the administration login for the management console, if so use the default user name and password; 'admin'/'changeit'.   Demonstration 4 (C# client) The final demonstration was the Visual Studio 2012 project using the Event Store client – referenced directly as C:\EventStore\v2.0.1\EventStore.ClientAPI.dll, although you can switch this to the latest Event Store client NuGet package. The source code provides a console app for viewing projections with the projection manager (HTTP connection), as well as containing a full set of data for the entire DDDNorth3 agenda.  It also deals with the strategy for reading newest events backwards to older events and ignoring older events that have been superseded. Resources Event Store home page: http://www.geteventstore.com/ Event Store source code on GitHub: https://github.com/eventstore/eventstore Event Store documentation on GitHub: https://github.com/eventstore/eventstore/wiki (includes index to @RobAshton’s blog series on Event Store at https://github.com/eventstore/eventstore/wiki#rob-ashton---projections-series) Event Store forum in Google Groups: https://groups.google.com/forum/?fromgroups#!forum/event-store TopShelf Windows service wrapper is available on github: https://gist.github.com/trbngr/5083266

    Read the article

  • Event ID for modified GPOs

    - by Hinek
    I have to know, who (usersid or loginname) changed a specified GPO for a specified OU in the Active Directory. Given our audit settings include this, what would be the right Event ID to look for?

    Read the article

  • Chester Devs Presentation and source code &ndash; &lsquo;Event Store - an introduction to a DSD for event sourcing and notifications&rsquo;

    - by Liam Westley
    Originally posted on: http://geekswithblogs.net/twickers/archive/2013/11/11/chester-devs-presentation-and-source-code-ndash-lsquoevent-store.aspxThank you everyone at Chester Devs Thanks to Fran Hoey and all the people from Chester Devs. It was a hard drive up and back but the enthusiasm of the audience, with some great questions does make it worthwhile. Presentation and source code My presentation, source code, Event Store runners and text files containing the various command line parameters used for curl is now available on GitHub; https://github.com/westleyl/ChesterDevs-EventStore. Don’t worry if you don’t have a GitHub account, you don’t need one, you can just click on the Download Zip button on the right hand menu to download all the files as a single ZIP file.  If all you want is the PowerPoint presentation, go to https://github.com/westleyl/ChesterDevs-EventStore/blob/master/Powerpoint/Huddle-EventStore.pptx, and click on the View Raw button. Downloading and installing Event Store and Tools Download Event Store http://download.geteventstore.com – I unzipped these files into C:\EventStore\v2.0.1 Download Curl from http://curl.haxx.se/download.html – I downloaded Win64 Generic (with SSL) and unzipped these files into C:\curl version 7.31.0 Running the tools I used in my presentation Demonstration 1 (running Event Store) You can use one of my Event Store runner command files to run the single node version of Event Store, using default ports of 2213 for HTTP and 1113  for TCP, and with a wildcard HTTP pattern.  Both take a single command line parameter to specify the location of the data and log files.  The runners assume the single node executable is located in C:\EventStore\v2.0.1, and will placed data files and logs beneath C:\EventStore\Data, i.e. RunEventStore.cmd TestData1 This will create data files in C:\EventStore\Data\TestData1\Data and log files in C:\EventStore\Data\TestData1\logs. If, when running Event Store you may see the following message, [03288,15,06:23:00.622] Failed to start http server Access is denied You will either need to run Event Store in an administrator console window, or you can use the netsh command to create a firewall permission to allow HTTP listening (this will need to be run, once, in an administrator console window), netsh http add urlacl url=http://*:2213/ user=liam You can always delete this later by running the delete; netsh http delete urlacl url=http://*:2213/ If you want to confirm that everything is running OK, open the management console in a browser by navigating to http://127.0.0.1:2213. If at any point you are asked for a user name and password use the default of ‘admin’/‘changeit’. Demonstration 2 (reading and adding data, curl) In my second demonstration I used curl directly from the console to read streams, write events and then read back those events. On GitHub I have included is a set of curl commands, CurlCommandLine.txt, and a sample data file, SampleData.json, to load an event into a DDDNorth3 stream. As there is not much data in the Event Store at this point I used the $stats-127.0.0.1:2113 which is a stream containing performance statistics for Event Store and is updated every 30 seconds (default). Demonstration 3 (projections) On GitHub I have included a sample projection, Projection-ByRoom.txt, which will create streams based on the room on which a session was held on the DDDNorth3 agenda. Browse to the management console, http://127.0.0.1:2213.  Click on Projections, New Projection, give it a name, Sessions-ByRoom, and copy in the JavaScript in the Projection-ByRoom.txt file.  Select Continuous, tick Emit Enabled and then click on Post. It should run immediately. You may by challenged for the administration login for the management console, if so use the default user name and password; 'admin'/'changeit'. Demonstration 4 (C# client) The final demonstration was the Visual Studio 2012 project using the Event Store client – referenced directly as C:\EventStore\v2.0.1\EventStore.ClientAPI.dll, although you can switch this to the latest Event Store client NuGet package. The source code provides a console app for viewing projections with the projection manager (HTTP connection), as well as containing a full set of data for the entire DDDNorth3 agenda.  It also deals with the strategy for reading newest events backwards to older events and ignoring older events that have been superseded. Resources Event Store home page: http://www.geteventstore.com/ Event Store source code on GitHub: https://github.com/eventstore/eventstore Event Store documentation on GitHub: https://github.com/eventstore/eventstore/wiki (includes index to @RobAshton’s blog series on Event Store at https://github.com/eventstore/eventstore/wiki#rob-ashton---projections-series) Event Store forum in Google Groups: https://groups.google.com/forum/?fromgroups#!forum/event-store TopShelf Windows service wrapper is available on github: https://gist.github.com/trbngr/5083266

    Read the article

  • x-dom-event-stream in Opera 10 Only Working on First Event

    - by Brad
    I have a python script (in the CherryPy framework) that sends Event: and data: text as this Opera blog post describes to a client browser. The javascript that recieves the x-dom-event-stream content is almost identical to what they show in the blog post. However, the browser displays only the first event sent. Anyone know what I'm missing? I tried a few older versions of Opera and found that it works in Opera 9.52 but not in any newer versions. What did they change? Here is the python code: class dumpData(object): def index(self): cherrypy.response.headers['Content-Type'] = "application/x-dom-event-stream" def yieldData(): i = 0 while 1: yield "Event: count\n" yield "data: " yield i yield "\n\n" i = i + 1 time.sleep(3); return yieldData() index._cp_config = {'response.stream': True} index.exposed = True And here is the javascript/html. Making a request to /data/ runs the python function above. <head> <script> onload = function() { document.getElementById("count").addEventListener("cout", cout, false); } function count(e) { document.getElementById("stream").firstChild.nodeValue = e.data; } </script> <event-source id="count" src="/data/"> </head> <body> <div id="stream"></div> </body> Opening the direct /data/ url in Firefox saves the stream to a file. So I know the output is in the correct format and that the stream works at all.

    Read the article

  • JavaScript or PHP based WYSIWYG vector based image editor

    - by Jeroen Pluimers
    For a PHP based site of a client, I'm looking for a vector based image editor that allows: end user creation of vectored images consisting of objects supports upload of bitmap images to be used as objects inside the vector image supports adding text objects to add to the vector image, and change properties (font name, font style, font size) of the text objects preferably supports layering or grouping of objects inside the vector image integrates nicely with a PHP based site (so a PHP or JavaScript library is preferred) can store the vector image in SVG, EPS or PDF Both commercial and FOSS solutions are OK. Any idea where to find such a library? --jeroen

    Read the article

  • We are moving an Access based corporate front-end into a Web-based App

    - by Max Vernon
    We have an enterprise application with a front end written in Microsoft Access 2003 that has evolved over the past 6 years. The back end data, and a fair amount of back-end logic is contained within several Microsoft SQL Server databases. This front end app consists of around 180 forms, and over 120,000 lines of code, and interacts with VB.Net DLLs that support various critical functions used by our sales force. The current system makes use of 3 monitors to display various information; the Access app uses COM+ to control Microsoft Outlook and Internet Explorer for various purposes. The Access front end sometimes occupies 2 screens, automatically resizing itself based on Windows API-reported screen dimensions. The app also uses a Google map to present data to our agents, and allows two-way interactivity with the map through COM+ connectivity to JavaScript contained in the Google map. At the urging of senior management, we are looking to completely rewrite this application using some web-based technology, such as ASP.Net or perhaps a LAMP stack (the thinking with the LAMP stack thing is "free" is pretty cheap). We want to move to a web-based app so we can eliminate the dependency on our physical location for hiring new sales force members. Currently, our main office is full to capacity, and we need to continue growing the company. Does anyone have any thoughts on what would be the best technology to use for a web-based app of this magnitude? Keeping in mind the app is dependent on back-end services on our existing infrastructure. The app handles financial data and personal customer data, among other things. [I've looked at Best practices for moving large MS Access application towards .Net? and read the answers, and most of the comments. Interesting reading, and has some valid points, but our C.O.O. and contracted Software Architect are pushing for a full web-based app, not a .Net Windows App]

    Read the article

  • How to add event receiver to SharePoint2010 content type programmatically

    - by ybbest
    Today , I’d like to show how to add event receiver to How to add event receiver to SharePoint2010 content type programmatically. 1. Create empty SharePoint Project and add a class called ItemContentTypeEventReceiver and make it inherit from SPItemEventReceiver and implement your logic as below public class ItemContentTypeEventReceiver : SPItemEventReceiver { private bool eventFiringEnabledStatus; public override void ItemAdded(SPItemEventProperties properties) { base.ItemAdded(properties); UpdateTitle(properties); } private void UpdateTitle(SPItemEventProperties properties) { SPListItem addedItem = properties.ListItem; string enteredTitle = addedItem["Title"] as string; addedItem["Title"] = enteredTitle + " Updated"; DisableItemEventsScope(); addedItem.Update(); EnableItemEventsScope(); } public override void ItemUpdated(SPItemEventProperties properties) { base.ItemUpdated(properties); UpdateTitle(properties); } private void DisableItemEventsScope() { eventFiringEnabledStatus = EventFiringEnabled; EventFiringEnabled = false; } private void EnableItemEventsScope() { eventFiringEnabledStatus = EventFiringEnabled; EventFiringEnabled = true; } } 2.Create a Site or Web(depending or your requirements) scoped feature and implement your feature event handler as below: public override void FeatureActivated(SPFeatureReceiverProperties properties) { SPWeb web = GetFeatureWeb(properties); //http://karinebosch.wordpress.com/walkthroughs/event-receivers-theory/ string assemblyName =  System.Reflection.Assembly.GetExecutingAssembly().FullName; const string className = "YBBEST.AddEventReceiverToContentType.ItemContentTypeEventReceiver"; SPContentType contentType= web.ContentTypes["Item"]; AddEventReceiverToContentType(className, contentType, assemblyName, SPEventReceiverType.ItemAdded, SPEventReceiverSynchronization.Asynchronous); AddEventReceiverToContentType(className, contentType, assemblyName, SPEventReceiverType.ItemUpdated, SPEventReceiverSynchronization.Asynchronous); contentType.Update(); } protected static void AddEventReceiverToContentType(string className, SPContentType contentType, string assemblyName, SPEventReceiverType eventReceiverType, SPEventReceiverSynchronization eventReceiverSynchronization) { if (className == null) throw new ArgumentNullException("className"); if (contentType == null) throw new ArgumentNullException("contentType"); if (assemblyName == null) throw new ArgumentNullException("assemblyName"); SPEventReceiverDefinition eventReceiver = contentType.EventReceivers.Add(); eventReceiver.Synchronization = eventReceiverSynchronization; eventReceiver.Type = eventReceiverType; eventReceiver.Assembly = assemblyName; eventReceiver.Class = className; eventReceiver.Update(); } 3.Deploy your solution and now you have a event receiver that attached to the Item contentType. You can download the complete source code here.You can also check how to add event receiver to a list using SharePoint event receiver item in Visual Studio2010 in my previous blog.

    Read the article

  • Push-Based Events in a Services Oriented Architecture

    - by Colin Morelli
    I have come to a point, in building a services oriented architecture (on top of Thrift), that I need to expose events and allow listeners. My initial thought was, "create an EventService" to handle publishing and subscribing to events. That EventService can use whatever implementation it desires to actually distribute the events. My client automatically round-robins service requests to available service hosts which are determined using Zookeeper-based service discovery. So, I'd probably use JMS inside of EventService mainly for the purpose of persisting messages (in the event that a service host for EventService goes down before it can distribute the message to all of the available listeners). When I started considering this, I began looking into the differences between Queues and Topics. Topics unfortunately won't work for me, because (at least for now), all listeners must receive the message (even if they were down at the time the event was pushed, or hadn't made a subscription yet because they haven't completed startup (during deployment, for example) - messages should be queued until the service is available). However, I don't want EventService to be responsible for handling all of the events. I don't think it should have the code to react to events inside of it. Each of the services should do what it needs with a given event. This would indicate that each service would need a JMS connection, which questions the value of having EventService at all (as the services could individually publish and subscribe to JMS directly). However, it also couples all of the services to JMS (when I'd rather that there be a single service that's responsible for determining how to distribute events). What I had thought was to publish an event to EventService, which pulls a configuration of listeners from some configuration source (database, flat file, irrelevant for now). It replicates the message and pushes each one back into a queue with information specific to that listener (so, if there are 3 listeners, 1 event would become 3 events in JMS). Then, another thread in EventService (which is replicated, running on multiple hots) would be pulling from the queue, attempting to make the service call to the "listener", and returning the message to the queue (if the service is down), or discarding the message (if the listener completed successfully). tl;dr If I have an EventService that is responsible for receiving events and delegating service calls to "event listeners," (which are really just endpoints on other services), how should it know how to craft the service call? Should I create a generic "Event" object that is shared among all services? Then, the EventService can just construct this object and pass it to the service call. Or is there a better answer to this problem entirely?

    Read the article

< Previous Page | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >