I recently started developing using akka event sourcing/cluster sharding, and thanks to the online resources I think I understood the basic concepts and how to create a simple application with it. I am however struggling to apply this methodology in a slightly more complex data structure:
As an example, let's think about webpages and URLs.
Each Page can be represented with an actor in the cluster (having its unique id as the path of the page, e.g. /questions/60037683).
On each page I can issue commands such as
Create page (so if the page does not exist, it will be created)
Edit page (editing the details of the page)
Get page content (and children)
Etc.
When issuing commands to single pages, everything is easy as it's "written on the manual". But I have the added the complexity that a web page can have children, so when creating a "child page" I need the parent to update references to its children.
I thought of some possible approaches, but they feel incomplete.
Sending all events to the single WebPage and when creating a page, finding the parent page (if any) and communicate that a new child has been added
Sending all events to the single WebPage, and when creating a page, the message is sent to the parent, and then it will create a new command that will tell the child to initialize
Creating an infrastructure as WebPageRepository that will keep track of the page tree and will relay CRUD commands to all web page actors.
My real problem is, I think, handling the return of Futures properly when relaying messages to other actors that have to actually perform the job.
I'm making a lot of confusion and some reading resources would be greatly appreciated.
Thanks for your time.
EDIT: the first version was talking about a generical hierarchical file-system-like structure. I updated with the real purpose, webpages and urls and tried to clarify better my issues
After some months of searching, I reached the conclusion that what I'm doing is trying to have actors behave transactionally, so that when something is created, the parent is also updated in a safe manner, meaning that if one operation fails, all operations who completed successfully are rolled back.
The best pattern for this, in my opinion, proved to be the saga pattern, which adds a bit of complexity to the whole process, but in the long run it does what I needed.
Basically I ended up implementing the main actor as a stand alone piece (as it should be) that can receive create commands and add children commands.
There is then a saga actor which takes care of creating the content, adding the child to the parent and rolling back everything if something fails during the process.
If someone else has a better solution, I'll be glad to hear it out
Related
Assume a page that shows a complex data structure (for example, an article with many details). This view will be reused from time to time by rebinding it to different articles.
Now, I noticed that the ODataModel keeps all used article entities in memory (also if they are no longer bound to any control).
This will lead to two issues:
Memory consumption increases over time (if application will not be reloaded).
If the application forces a refresh of the data model, all entities will be loaded (also not used).
The second issue seems to be the bigger problem. It slows down the speed of the application.
I have not found a solution for that problem. If I use refresh(true, true) it seems all data will be reloaded.
Is there an way to clean the model?
Edit
Lets say you have a list of thousands of articles. User can click on one of the articles and will navigate to a detailed screen of that article.
The OData model in client side will cache this. To see it, do something like:
var oModel = this.getModel("modelName");
look with the debugger into oModel.oData.
If the user now navigates back and chooses the next article, this will be cached as well.
If user does this 1000 times, all articles are now in the model.
If you trigger a oModel.refresh(true);, all these data (of 1000 articles) will be reloaded not only the one bound to the view.
Now my application is not about showing article information. It's a more complex structure with subitems. Each time user is visiting this page, more data will be cached (and re-fetched in case of a refresh call on the model).
Edit 2
The function updateBindings(bForceUpdate?) seems to help a little bit.
Anyhow, the data accumulation is still there in the ODataModel class.
That means: Each visited data path will stay in memory since the next reload (F5) of the full page. If someone uses such an application over a day, the data accumulates and a refresh call on the model will read all data again, if still bound to a view or not.
Try deleteCreatedEntry(oContext). Even though this is not the supposed use case for this method it might work to delete an entity from the model without triggering a backend request.
You could also try if updateBindings(bForceUpdate?) only triggers an update on actually bound entities.
1) I do not really understand your problem here. What is it exactly that you do? OData always holds the result of your request plus a queue of changes to that request. If you create lots of entries while your application is running, of course the memory consumption will increase. If you want to revert back to the original request you can use resetChanges(). THis way the used memory should decrease again. But you lose all your changes to the model.
2) Maybe you should look into Odata filtering (http://www.odata.org/getting-started/basic-tutorial/) so that you only load the entities you really want. If you only want a part of the entity loaded then you should maybe redesign your entities to avoid a lot of overhead.
It is hard to speculate what your exact problem is.
Well, if you know exactly what are you doing, you can try something like this:
this.getModel("modelname").aBindings = []
Better solution would be go through the aBindings array and remove redundant bindings.
I'm building a simple REST API for generating some objects that must be created and sent periodically out of the API. The nature of the objects doesn't matter, neither the framework supporting the REST interface (Spray, Play Framework, whatever else). My question is, what would be a good scalable actor design for this system using Akka? Suppose the service crashes or it's migrated or whatever that causes to stop it. In order to recover the description of the tasks about what objects must be sent and when, is akka-persistence a good way to go here? or it's better to persist such things in a traditional DB?
Thanks.
NOTE: also I would like to know, supposing there's some actor which is not stateful himself, but creates many children actors, if it's a good practice to use akka-persistence in order to replay the messages which causes this actor to create his children again (the children being also non-stateful).
In a traditional DB you would most likely end up modeling this with timestamps and events, and with event sourcing this is already the native model.
Akka-persistence would be a natural fit for this scenario since it will persist every event about what objects must be created and sent periodically out. The snapshot support will also help with speed of recovery when the number of events gets very large.
In the case of crashes or migration, the recovery process will handle this just fine.
Regarding your note, if the actor is truly stateless then there is no need to persist the events that cause the children to be created since they can be recreated on demand. If the existence of the children does need to be recovered, then the actor is not stateless. In that case then it may indeed make sense to persist those events.
I´m just trying to wrap my head around CQRS(/ES). I have not done anything serious with CQRS. Probably I´m just missing something very fundamental right now. Currently, I´m reading "Exploring CQRS and Event Sourcing". There is one sentence that somehow puzzles me in regards to commands:
"A single recipient processes a command."
I´ve seen this also in the CQRS sample application from Greg Young (FakeBus.cs) where an exception is thrown when more then one command handler is registered for any command type.
For me, this is an indication that this is a fundamental principle for CQRS (or Commands?). What is the reason? For me, it is somewhat counter-intuitive.
Imagine I have two components that need to perform some action in response to a command (it doesn´t matter if I have two instances of the same component or two independent components). Then I would need to create a handler that delegates the command to these components.
In my opinion, this is introducing an unnecessary dependency. In terms of CQRS, a command is nothing more than a message that is sent. I don´t get the reason why there should be only one handler for this message.
Can someone tell me what I am missing here? There is probably a very good reason for this that I just don´t see right now.
Regards
I am by no means an expert myself with CQRS, but perhaps I can help shed some light.
"A single recipient processes a command.", What is the reason?
One of the fundamental reasons for this is transactional consistency. A command needs to be handled in one discrete (and isolated) part of the application so that it can be committed in a single transaction. As soon as you start to have multiple handlers, distributing the application beyond a single process (and maintaining transactional consistency) is nearly impossible. So, while you could design that way, it is not recommended.
Hope this helps.
Imagine I have two components that need to perform some action in response to a command (it doesn´t matter if I have two instances of the same component or two independent components). Then I would need to create a handler that delegates the command to these components.
That's the responsibility of events.
A command must be handled by one command handler and must change the state for a single aggregate root. The aggregate root then raises one or more events indicating that something happened. These events can have multiple listeners that perform desired actions.
For example, you have a PurchaseGift command. Your command handler loads the Purchase aggregate root and performs the desired operation raising a GiftPurchased event. You can have one or more listeners to the GiftPurchase event, one for sending an email to the buyer confirming the operation and another to send the gift by mail.
I am attempting to learn and apply the CQRS design approach (pattern and architecture) to a new project but seem to be missing a key piece.
My client application executes a query and retrieves a list of light-weight, read-only DTOs from the read model. The user selects an item and clicks a button to initiate some action. The action is performed by creating and sending the corresponding command object to the write model (where the command handler carries out the action, updates the data store, etc.) At some point, however, I need to update the UI to reflect changes to the state of the application resulting from the action.
How does the UI know when it is time to refresh the original list?
Additional Info
I have noticed that most articles/blogs discussing CQRS use MVC client apps in their examples. I am working on a Silverlight client right now and am beginning to wonder if the pattern simply doesn't work in that case.
Follow-Up Question
After thinking more about Bartlomiej's response and subsequent discussion, I am wondering about error handling in CQRS. Given that commands are basically fire-and-forget asynchronous operations, how do we report an error condition to the UI?
I see 'refreshing the UI' to take one of two forms:
The operation succeeds, data has changed and the UI should be updated to reflect these changes
The operation fails, data has not changed but the user should be notified of the failure and potential corrective actions.
Even with a Post-Redirect-Get pattern in an MVC, you can't really Redirect until you know the outcome of the operation. None of the examples I've seen thus far address these real-world concerns.
I've been struggling with similar issues for a WPF client. The re-query trigger for any data is dependent on the data your updating, commands tend to fall into categories:
The command is a true fire and forget method, it informs the back-end of a state change but this change does not need to be reflected in the UI, or the change simply isn't important to the UI.
The command will alter the result of a single query
The command will alter the result of multiple queries, usually (in my domain at least) in a cascading fashion, that is, changing the state of a single "high level" piece of data will likely affect many "low level" caches.
My first trigger is the page load, very few items are exempt from this as most pages must assume data has been updated since it was last visited. Though some systems may be able to escape with only updating financial and other critical data in this way.
For short commands I also update data when 'success' is returned from a command. Though this is mostly laziness as IMHO all CQRS commands should be fired asynchronously. It's still an option I couldn't live without but one you may have to if your implementation expects high latency between command and query.
One pattern I'm starting to make use of is the mediator (most MVVM frameworks come with one). When I fire a command, I also fire a message to the mediator specifying which command was launched. Each Cache (A view model property Retriever<T>) listens for commands which affect it and then updates appropriately. I try to minimise the number of messages while still minimising the number of caches that update unnecessary from a single message so I'll (hopefully) eventually end up with a shortlist of update reasons, with each 'reason' updating a list of caches.
Another approach is simple honesty, I find that by exposing graphically how the system updates itself makes users more willing to be patient with it. On firing a command show some UI indicating you're waiting for the successful response, on error you could offer to retry / show the error, on success you start the update of the relevant fields. Baring in mind that this command could have been fired from another terminal (of which you have no knowledge) so data will need to timeout eventually to avoid missing state changes invoked by other machines also.
Noting the irony that the only efficient method of updating cache's and values on a client is to un-separate the commands and queries again, be it through hardcoding or something like a hashmap.
My two cents.
I think MVVM actually fits into CQRS quite well. The ViewModel simply becomes an observable ReadModel.
1 - You initialize your ViewModel state via a query on the ReadModel.
2 - Changes on your ViewModel are automatically reflected on any Views that are bound to it.
3 - Certain changes on your ViewModel trigger a command to propegate to a message queue, an object responsible for sending those commands to the server takes those messages off the queue and sends them to the WriteModel.
4 - Clients should be well formed, meaning the ViewModel should have performed appropriate validation before it ever triggered the command. Once the command has been triggered, any event notifications can be published onto an event bus for the client to communicate changes to other ViewModels or components in the system interested in those changes. These events should carry the relevant information necessary. Typically, this means that other view models usually don't have to re-query the read model as a result of the change unless they are dependent on other data that needs to be retrieved.
5 - There is an object that connects to the message bus on the server for real-time push notifications when other clients make changes that this client is interested in knowing about, falling back to long-polling if necessary. It propagates those to the internal message bus that ties the components on the client together.
6 - The last part to handle is the fact that clients can be occasionally connected, which should be the only reason a command fails (they don't have internet access at the moment), which is when the client should be notified of problems.
In my ASP.NET MVC 3 I use 2 techniques depending on use case:
already well-known Post-Redirect-Get pattern which fits nicely with CQRS. Your MVC action that triggers the command returns a redirection to action that performs a query.
in some cases, like real-time updates of other clients, I rely on domain events/messages. I create an event handler that uses singlarR to push changes to all connected and interested clients.
There are two major ways you can take as far as I know :
1) design your UI , so that the user does not see its changes right away. Like for instance a message to tell him his action is a success, and offering him different choices to continue his work. this should buy you enough time to have updated your readmodel.
2) more complex, but you might keep the information you have send to the server and shows them in the interface.
The most important I guess, educate your user if you can so that they know why the data is not here... yet!
I am thinking about it only now, but these are for sync command handling, not async, in async things go really harder on the brain...the client interface becomes an event eater too..
I'm building an app that deals with customer queries, where I want to route the query through a decision tree showing appropriate views before taking some automated action against their query. Kind of like the game "20 questions"! Based on the answers at each stage, the path through the app will change.
I was thinking of using MVC, because there are only a few "types" of route and outcome - so I could build fewer pages that way, one to handle each type rather than one for each step. I was also thinking of using Workflow 4 to manage the page flow, because the flowchart model maps pretty nicely to what I'm trying to do.
Does anyone know any good reference apps that use Workflow for this kind of thing?
Thanks
Richard
There where a number of examples using WF3 doing this sort of thing but I haven't seen any for WF4. I suppose it is possible to do but it means running the workflow synchronously and checking the bookmarks as soon as it becomes idle to see which operations are enabled at the moment. That should be possible using a custom SynchronizationContext that does things synchronous and using the Idle callback on the WorklfowApplication to check the current bookmarks.
I actually went with a different option in the end - I wrote a "GetNextAction" function that returned an ActionResult object based on my flowchart logic and state of objects. The controller processes whatever form inputs it's received, updates the object, then calls GetNextAction and returns the result of that function. Seems to be working out ok!