Im learning WebFlux.
Wiki says that reactive programming is:
For example, in an imperative programming setting, a:=b+c would mean that a is being assigned the result of b+c in the instant the expression is evaluated, and later, the values of b and/or c can be changed with no effect on the value of a.
However, in reactive programming, the value of a is
automatically updated whenever the values of b and/or c change; without the program having to re-execute the sentence
a:=b+c to determine the presently assigned value of a.
Ok. When Im reproducing example like:
#RestController
public class PersonController {
private final PersonRepository repository;
public PersonController(PersonRepository repository) {
this.repository = repository;
}
#PostMapping("/person")
Mono<Void> create(#RequestBody Publisher<Person> personStream) {
return this.repository.save(personStream).then();
}
#GetMapping("/person")
Flux<Person> list() {
return this.repository.findAll();
}
#GetMapping("/person/{id}")
Mono<Person> findById(#PathVariable String id) {
return this.repository.findOne(id);
}
}
I'm Posting 2 persons. (on the chrome page 1)
Then getting list of all persons (on the chrome page 2)
Then adding one more person (on the chrome page 3)
Then I'm getting back to the page 2 (with no refreshing), I dont see updated list of persons, should I?
Also, how should work UPDATE/DELETE operations here?
I guess you're referring to the reactive programming wikipedia page and maybe reading too much into that example.
This example (and the famous spreadsheet one) usually point to UI rich applications that are listening to user events and publishing application events to update the UI.
Reactive programming and Reactive Streams by themselves aren't enough to set up such an infrastructure.
In your Controller, operations are performed and values are published in a reactive way: with backpressure support and access to a reactive API to compose them. Once the JSON response is rendered, the client doesn't receive new elements from the server.
You can create such a system though, by publishing events and having a persistent connection (SSE, for example) between server and browser.
Related
I'm looking for a solution that will have the backend publish an event to the frontend as soon as a modification is done on the server side. To be more concise I want to emit a new List of objects as soon as one item is modified.
I've tried implementing on a SpringBoot project, that uses Reactive Web, MongoDB which has a #Tailable cursor that publish an event as soon as the capped collection is modified. The problem is that the capped collection has some limitation and is not really compatible with what I want to do. The thing is I cannot update an existing element if the new one has a different size(as I understood this is illegal because you cannot make a rollback).
I honestly don't even know if it's doable, but maybe I'm lucky and I'll run into a rocket scientist right here that will prove otherwise.
Thanks in advance!!
*** EDIT:
Sorry for the vague question. Yes I'm more focused on the HOW, using the Spring Reactive framework.
When I had a similar need - to inform frontend that something is done on the backend side - I have used a message queue.
I have published a message to the queue from the backend and the frontend consumed the message.
But I am not sure if that is what you're looking for.
if you are using webflux with spring reactor, I think you can simply have a client request with content-type as 'text/event-stream' or 'application/stream+json' and You shall have API that can produce those content-type. This gives you SSE model without too much effort.
#GetMapping(value = "/stream", produces = {MediaType.TEXT_EVENT_STREAM_VALUE, MediaType.APPLICATION_STREAM_JSON_VALUE, MediaType.APPLICATION_JSON_UTF8_VALUE})
public Flux<Message> get(HttpServletRequest request) {
Just as an idea - maybe you need to use a web socket technology here:
The frontend side (I assume its a client side application that runs in a browser, written in react, angular or something like that) can establish a web-socket communication with the backend server.
When the process on backend finishes, the message from backend to frontend can be sent.
You can do emitting changes by hand. For example:
endpoint:
public final Sinks.Many<SimpleInfoEvent> infoEventSink = Sinks.many().multicast().onBackpressureBuffer();
#RequestMapping(path = "/sseApproach", produces = MediaType.TEXT_EVENT_STREAM_VALUE)
public Flux<ServerSentEvent<SimpleInfoEvent>> sse() {
return infoEventSink.asFlux()
.map(e -> ServerSentEvent.builder(e)
.id(counter.incrementAndGet() + "")
.event(e.getClass().getName())
.build());
}
Code anywhere for emitting data:
infoEventSink.tryEmitNext(new SimpleInfoEvent("any custom event"));
Watch out of threads and things like "subscribeOn", "publishOn", but basically (when not using any third party code), this should work good enough.
The question assumes the use of Event Sourcing.
When rebuilding current state by replaying events, event handlers should be idempotent. For example, when a user successfully updates their username, a UsernameUpdated event might be emitted, the event containing a newUsername string property. When rebuilding current state, the appropriate event handler receives the UsernameUpdated event and sets the username property on the User object to the newUsername property of the UsernameUpdated event object. In other words, the handling of the same message multiple times always yields the same result.
However, how does such an event handler work when integrating with external services? For example, if the user wants to reset their password, the User object might emit a PasswordResetRequested event, which is handled by a portion of code that issues a 3rd party with a command to send an SMS. Now when the application is rebuilt, we do NOT want to re-send this SMS. How is this situation best avoided?
There are two messages involved in the interaction: commands and events.
I do not regard the system messages in a messaging infrastructure the same as domain events. Command message handling should be idempotent. Event handlers typically would not need to be.
In your scenario I could tell the aggregate root 100 times to update the user name:
public UserNameChanged ChangeUserName(string username, IServiceBus serviceBus)
{
if (_username.Equals(username))
{
return null;
}
serviceBus.Send(new SendEMailCommand(*data*));
return On(new UserNameChanged{ Username = userName});
}
public UserNameChanged On(UserNameChanged #event)
{
_username = #event.UserName;
return #event;
}
The above code would result in a single event so reconstituting it would not produce any duplicate processing. Even if we had 100 UserNameChanged events the result would still be the same as the On method does not perform any processing. I guess the point to remember is that the command side does all the real work and the event side is used only to change the state of the object.
The above isn't necessarily how I would implement the messaging but it does demonstrate the concept.
I think you are mixing two separate concepts here. The first is reconstructing an object where the handlers are all internal methods of the entity itself. Sample code from Axon framework
public class MyAggregateRoot extends AbstractAnnotatedAggregateRoot {
#AggregateIdentifier
private String aggregateIdentifier;
private String someProperty;
public MyAggregateRoot(String id) {
apply(new MyAggregateCreatedEvent(id));
}
// constructor needed for reconstruction
protected MyAggregateRoot() {
}
#EventSourcingHandler
private void handleMyAggregateCreatedEvent(MyAggregateCreatedEvent event) {
// make sure identifier is always initialized properly
this.aggregateIdentifier = event.getMyAggregateIdentifier();
// do something with someProperty
}
}
Surely you wouldn't put code that talks to an external API inside an aggregate's method.
The second is replaying events on a bounded context which could cause the problem you are talking about and depending on your case you may need to divide your event handlers into clusters.
See Axon frameworks documentation for this point to get a better understanding of the problem and the solution they went with.
Replaying Events on a Cluster
TLDR; store the SMS identifier within the event itself.
A core principle of event sourcing is "idempotency". Events are idempotent, meaning that processing them multiple times will have the same result as if they were processed once. Commands are "non-idempotent", meaning that the re-execution of a command may have a different result for each execution.
The fact that aggregates are identified by UUID (with a very low percentage of duplication) means that the client can generate the UUIDs of newly created aggregates. Process managers (a.k.a., "Sagas") coordinate actions across multiple aggregates by listening to events in order to issue commands, so in this sense, the process manager is also a "client". Cecause the process manager issues commands, it cannot be considered "idempotent".
One solution I came up with is to include the UUID of the soon-to-be-created SMS in the PasswordResetRequested event. This allows the process manager to only create the SMS if it does not yet already exist, hence achieving idempotency.
Sample code below (C++ pseudo-code):
// The event indicating a password reset was successfully requested.
class PasswordResetRequested : public Event {
public:
PasswordResetRequested(const Uuid& userUuid, const Uuid& smsUuid, const std::string& passwordResetCode);
const Uuid userUuid;
const Uuid smsUuid;
const std::string passwordResetCode;
};
// The user aggregate root.
class User {
public:
PasswordResetRequested requestPasswordReset() {
// Realistically, the password reset functionality would have it's own class
// with functionality like checking request timestamps, generationg of the random
// code, etc.
Uuid smsUuid = Uuid::random();
passwordResetCode_ = generateRandomString();
return PasswordResetRequested(userUuid_, smsUuid, passwordResetCode_);
}
private:
Uuid userUuid_;
string passwordResetCode_;
};
// The process manager (aka, "saga") for handling password resets.
class PasswordResetProcessManager {
public:
void on(const PasswordResetRequested& event) {
if (!smsRepository_.hasSms(event.smsUuid)) {
smsRepository_.queueSms(event.smsUuid, "Your password reset code is: " + event.passwordResetCode);
}
}
};
There are a few things to note about the above solution:
Firstly, while there is a (very) low possibility that the SMS UUIDs can conflict, it can actually happen, which could cause several issues.
Communication with the external service is prevented. For example, if user "bob" requests a password reset that generates an SMS UUID of "1234", then (perhaps 2 years later) user "frank" requests a password reset that generates the same SMS UUID of "1234", the process manager will not queue the SMS because it thinks it already exists, so frank will never see it.
Incorrect reporting in the read model. Because there is a duplicate UUID, the read side may display the SMS sent to "bob" when "frank" is viewing the list of SMSes the system sent him. If the duplicate UUIDs were generated in quick succession, it is possible that "frank" would be able to reset "bob"s password.
Secondly, moving the SMS UUID generation into the event means you must make the User aggregate aware of the PasswordResetProcessManager's functionality (but not the PasswordResetManager itself), which increases coupling. However, the coupling here is loose, in that the User is unaware of how to queue an SMS, only that an SMS should be queued. If the User class were to send the SMS itself, you could run into the situation in which the SmsQueued event is stored while the PasswordResetRequested event is not, meaning that the user will receive an SMS but the generated password reset code was not saved on the user, and so entering the code will not reset the password.
Thirdly, if a PasswordResetRequested event is generated but the system crashes before the PasswordResetProcessManager can create the SMS, then the SMS will eventually be sent, but only when the PasswordResetRequested event is re-played (which might be a long time in the future). E.g., the "eventual" part of eventual consistency could be a long time away.
The above approach works (and I can see that it should also work in more complicated scenarious, like the OrderProcessManager described here: https://msdn.microsoft.com/en-us/library/jj591569.aspx). However, I am very keen to hear what other people think about this approach.
Using JAX-WS 2, I see an issue that others have spoken about as well. The issue is that if a SOAP message is received inside a handler, and that SOAP message is large - whether due to inline SOAP body elements that happen to have lots of content, or due to MTOM attachments - then it is dangerously easy to get an OutOfMemoryError.
The reason is that the call to getMessage() seems to set off a chain of events that involve reading the entire SOAP message on the wire, and creating an object (or objects) representing what was on the wire.
For example:
...
public boolean handleMessage(SOAPMessageContext context)
{
// for a large message, this will cause an OutOfMemoryError
System.out.println( context.getMessage().countAttachments() );
...
My question is: is there a known mechanism/workaround for dealing with this? Specifically, it would be nice to access the SOAP part in a SOAP message without forcing the attachments (if MTOM for example) to also be vacuumed up.
For those who run their app on JBoss 6 & 7 (with Apache CXF)... I was able to troubleshoot the problem by implementing my handler from the LogicalHandler interface instead of the SOAPHandler.
In this case your handleMessage() method would get the LogicalMessageContext context (instead of SOAPMessageContext) in the arguments that has no issues with the context.getMessage() call
There's actually a JAX-WS RI (aka Metro) specific solution for this which is very effective.
See https://javaee.github.io/metro/doc/user-guide/ch02.html#efficient-handlers-in-jax-ws-ri. Unfortunately that link is now broken but you can find it on WayBack Machine. I'll give the highlights below:
The Metro folks back in 2007 introduced an additional handler type, MessageHandler<MessageHandlerContext>, which is proprietary to Metro. It is far more efficient than SOAPHandler<SOAPMessageContext> as it doesn't try to do in-memory DOM representation.
Here's the crucial text from the original blog article:
MessageHandler:
Utilizing the extensible Handler framework provided by JAX-WS
Specification and the better Message abstraction in RI, we introduced
a new handler called MessageHandler to extend your Web Service
applications. MessageHandler is similar to SOAPHandler, except that
implementations of it gets access to MessageHandlerContext (an
extension of MessageContext). Through MessageHandlerContext one can
access the Message and process it using the Message API. As I put in
the title of the blog, this handler lets you work on Message, which
provides efficient ways to access/process the message not just a DOM
based message. The programming model of the handlers is same and the
Message handlers can be mixed with standard Logical and SOAP handlers.
I have added a sample in JAX-WS RI 2.1.3 showing the use of
MessageHandler to log messages and here is a snippet from the sample:
public class LoggingHandler implements MessageHandler<MessageHandlerContext> {
public boolean handleMessage(MessageHandlerContext mhc) {
Message m = mhc.getMessage().copy();
XMLStreamWriter writer = XMLStreamWriterFactory.create(System.out);
try {
m.writeTo(writer);
} catch (XMLStreamException e) {
e.printStackTrace();
return false;
}
return true;
}
public boolean handleFault(MessageHandlerContext mhc) {
.....
return true;
}
public void close(MessageContext messageContext) { }
public Set getHeaders() {
return null;
}
}
(end quote from 2007 blog post)
You can find a full example in the Metro GitHub repo.
What JAX-WS implementation runtime are you using? If there's a way to do this using the runtime built into WebSphere I'm certain there's a way to do this cleanly in other runtimes like Axis2 (proper), Apache CXF, and Metro/RI.
I am using the other way to reduce the memory costing, which is Message Accessor.
Instead of using context.getMessage(), I changed it to this way:
Object accessor = context.get("jaxws.message.accessor");
if (accessor != null) {
baosInString = accessor.toString();
}
Base on advice from IBM website. http://www-01.ibm.com/support/docview.wss?uid=swg1PM21151
What is the proper way to use Messenger class ?
I know it can be used for ViewModels/Views communications, but is it a good approach to use it in for a technical/business service layer ?
For example, a logging/navigation service registers for some messages in the constructors and is aware when these messages occurs in the app. The sender (ViewModel ou Service) does not reference the service interface but only messenger for sending messages. Here is a sample service :
using System;
using System.Windows;
using System.Windows.Navigation;
using Microsoft.Phone.Controls;
using App.Service.Interfaces;
using GalaSoft.MvvmLight.Messaging;
namespace App.Service
{
public class NavigationService : INavigationService
{
private PhoneApplicationFrame _mainFrame;
public event NavigatingCancelEventHandler Navigating;
public NavigationService()
{
Messenger.Default.Register<NotificationMessage<Uri>>(this, m => { this.NavigateTo(m.Content); });
}
public void NavigateTo(Uri pageUri)
{
if (EnsureMainFrame())
{
_mainFrame.Navigate(pageUri);
}
}
public void GoBack()
{
if (EnsureMainFrame()
&& _mainFrame.CanGoBack)
{
_mainFrame.GoBack();
}
}
private bool EnsureMainFrame()
{
if (_mainFrame != null)
{
return true;
}
_mainFrame = Application.Current.RootVisual as PhoneApplicationFrame;
if (_mainFrame != null)
{
// Could be null if the app runs inside a design tool
_mainFrame.Navigating += (s, e) =>
{
if (Navigating != null)
{
Navigating(s, e);
}
};
return true;
}
return false;
}
}
}
For me, the main use of a messenger is because it allows for communication between viewModels. Lets say you have a viewmodel that is used to provide business logic to a search function and 3 viewmodels on your page/window that want to process the search to show output, the messenger would be the ideal way to do this in a loosely-bound way.
The viewmodel that gets the search data would simply send a "search" message that would be consumed by anything that was currently registered to consume the message.
The benefits here are:
easy communication between viewmodels without each viewmodel having to know about each other
I can swap out the producer without affecting a consumer.
I can add more message consumers with little effort.
It keeps the viewmodels simple
Edit:
So, what about services?
ViewModels are all about how to present data to the UI. They take your data and shape it into something that can be presented to your View. ViewModels get their data from services.
A service provides the data and/or business logic to the ViewModel. The services job is to service business model requests. If a service needs to communicate/use other services to do its job these should be injected into the service using dependency injection. Services would not normally communicate with each other using a messenger. The messenger is very much about horizontal communication at the viewmodel level.
One thing I have seen done is to use a messenger as a mediator, where instead of injecting the service directly into a viewmodel the messenger is injected into the viewmodel instead. The viewmodel subscribes to an event and receives events containing models from the event. This is great if you're receiving a steady flow of updates or you're receiving updates from multiple services that you want to merge into a single stream.
Using a messenger instead of injecting a service when you're doing request/response type requests doesn't make any sense as you'll have to write more code to do this that you'd have to write just injecting the service directly and it makes the code hard to read.
Looking at your code, above. Imagine if you had to write an event for each method on there (Navigate, CanNavigate, GoBack, GoForward, etc). You'd end up with a lot of messages. Your code would also be harder to follow.
The preamble
We're implementing a MVC2 site that needs to consume an external API via https (We cannot use WCF or even old-style SOAP WebServices, I'm afraid). We're using AsyncController wherever we need to communicate with the API, and everything is running fine so far.
Some scenarios have come up where we need to make multiple API calls in series, using results from one step to perform the next.
The general pattern (simplified for demonstration purposes) so far is as follows:
public class WhateverController : AsyncController
{
public void DoStuffAsync(DoStuffModel data)
{
AsyncManager.OutstandingOperations.Increment();
var apiUri = API.getCorrectServiceUri();
var req = new WebClient();
req.DownloadStringCompleted += (sender, e) =>
{
AsyncManager.Parameters["result"] = e.Result;
AsyncManager.OutstandingOperations.Decrement();
};
req.DownloadStringAsync(apiUri);
}
public ActionResult DoStuffCompleted(string result)
{
return View(result);
}
}
We have several Actions that need to perform API calls in parallel working just fine already; we just perform multiple requests, and ensure that we increment AsyncManager.OutstandingOperations correctly.
The scenario
To perform multiple API service requests in series, we presently are calling the next step within the event handler for the first request's DownloadStringCompleted. eg,
req.DownloadStringCompleted += (sender, e) =>
{
AsyncManager.Parameters["step1"] = e.Result;
OtherActionAsync(e.Result);
AsyncManager.OutstandingOperations.Decrement();
}
where OtherActionAsync is another action defined in this same controller following the same pattern as defined above.
The question
Can calling other async actions from within the event handler cause a possible race when accessing values within AsyncManager?
I tried looking around MSDN but all of the commentary about AsyncManager.Sync() was regarding the BeginMethod/EndMethod pattern with IAsyncCallback. In that scenario, the documentation warns about potential race conditions.
We don't need to actually call another action within the controller, if that is off-putting to you. The code to build another WebClient and call .DownloadStringAsync() on that could just as easily be placed within the event handler of the first request. I have just shown it like that here to make it slightly easier to read.
Hopefully that makes sense! If not, please leave a comment and I'll attempt to clarify anything you like.
Thanks!
It turns out the answer is "No".
(for future reference incase anyone comes across this question via a search)