I am using JavaEE 7 and Glassfish 4 as my project setup and I would like to use websockets. As browser I have used Chrome 35.0.1916.153. My endpoint should receive a message through websockets and broadcast this message to other users within the same group.
If I open two browser tabs which open websocket connections to different groups [e.g. first 1, then 2 (the order matters!)] everything works as expected. But when I call the broadcast method by the first tab (groupId=1) the groupId path parameter has the value 2. It looks like it is the last onOpen call's value.
In this call the Session object looks like this:
session = (org.glassfish.tyrus.core.SessionImpl) SessionImpl{
uri=/test/group/1, <--
id='0fcff8ae-7ff8-42e8-a9e9-03cbb1f62562',
endpoint=EndpointWrapper{
endpointClass=null,
endpoint=org.glassfish.tyrus.core.AnnotatedEndpoint#59d91963,
uri='/test/group/2', <--
contextPath='/test'
}
}
The URI from session and endpoint differ which may result in the unwanted behavior.
Has anyone noted this yet?
My endpoint class looks like this:
#ServerEndpoint(value = "/group/{groupId}",
encoders = {MessageEncoder.class},
decoders = {MessageDecoder.class})
public class GroupEndpoint {
private Map<String, Set<Session>> groups = GroupSingleton.getGroups();
#OnMessage
public void broadcast(#PathParam("groupId") String groupId,
Message message,
Session session)
throws IOException, EncodeException {
for (Session session : groups.get(groupId))
session.getBasicRemote().sendObject(message);
}
#OnOpen
public void onOpen(#PathParam("groupId") String groupId,
Session session){
if(groups.containsKey(property)){
// then add peer to the list of viewers
groups.get(groupId).add(session);
}else{
// if not start a new list
Set<Session> sessions = Collections.synchronizedSet(new HashSet<Session>());
sessions.add(session);
groups.put(groupId, sessions);
}
}
#OnClose
public void onClose(#PathParam("groupId") String groupId, Session session) {
groups.get(groupId).remove(session);
if(groups.get(groupId).isEmpty())
groups.remove(groupId);
}
}
This is a known bug, tracked as TYRUS-203, which was fixed in Tyrus 1.1
Can you please useone of promoted Glassfish builds?
Thanks,
Pavel
Related
I have been developing some program with Spring and Redis but I am in stuck :(
What I want
when http request comes to server, then setex myKey, value and ttl.
when myKey is expired, then expired event comes to server and I will make a log.
Source code
#Configuration
public class RedisListenerContainerConfig {
#Bean("redisMyListenerContainer")
public RedisMessageListenerContainer redisLPRInOutCarExpiryListenerContainer(
RedisConnectionFactory redisConnectionFactory,
RedisLPRInOutCarExpiryListener redisLPRInOutCarExpiryListener) {
redisConnectionFactory.getConnection().setConfig("notify-keyspace-events", "Kx");
RedisMessageListenerContainer container = new RedisMessageListenerContainer();
container.setConnectionFactory(redisConnectionFactory);
container.addMessageListener(myListener, new PatternTopic("__keyspace#0__:myKey*"));
container.setErrorHandler(e -> log.error("error"));
return container;
}
}
// here expired event will be got
#Component
public class MyListener implements MessageListener {
#Override
public void onMessage(Message message, byte[] pattern) {
// to make log
}
}
#Service
public class MyService {
...
// here setex is generated
private void setExMyKey(String myKey) {
String value = "value";
redisTemplate.opsForValue().set(myKey, value, 60 * 10 ,TimeUnit.SECONDS);
final String listKey = "list_" + myKey;
if (redisTemplate.opsForList().size(listKey) < 1) {
redisTemplate.opsForList().leftPush(listKey, value);
}
}
}
info of Redis
standalone
keys: 610000
expires: 600000
Problem
Actually, the http requests keep coming to the server so it supposes to the ttl of mykey will never be zero.
But after taking a while, although ttl still remains, expired event is generated so many times.
127.0.0.1:6379> ttl myKey_3378
(integer) 1783
the suspicious part
According to Redis documents, I think, my function for setex does not change the value but only ttl.
And that could mean myKey is not really modified.
IMPORTANT all the commands generate events only if the target key is really modified.
For instance an SREM deleting a non-existing element from a Set will not actually change the value of the key, so no event will be generated.
Is it possible to be a issue of Redis lock?
My language is not English so I am sorry in advance not to understand very well.
Hopefully I solve this problem soon.
Thank you so much.
I have a CXF client configured in my Spring Boot app like so:
#Bean
public ConsumerSupportService consumerSupportService() {
JaxWsProxyFactoryBean jaxWsProxyFactoryBean = new JaxWsProxyFactoryBean();
jaxWsProxyFactoryBean.setServiceClass(ConsumerSupportService.class);
jaxWsProxyFactoryBean.setAddress("https://www.someservice.com/service?wsdl");
jaxWsProxyFactoryBean.setBindingId(SOAPBinding.SOAP12HTTP_BINDING);
WSAddressingFeature wsAddressingFeature = new WSAddressingFeature();
wsAddressingFeature.setAddressingRequired(true);
jaxWsProxyFactoryBean.getFeatures().add(wsAddressingFeature);
ConsumerSupportService service = (ConsumerSupportService) jaxWsProxyFactoryBean.create();
Client client = ClientProxy.getClient(service);
AddressingProperties addressingProperties = new AddressingProperties();
AttributedURIType to = new AttributedURIType();
to.setValue(applicationProperties.getWex().getServices().getConsumersupport().getTo());
addressingProperties.setTo(to);
AttributedURIType action = new AttributedURIType();
action.setValue("http://serviceaction/SearchConsumer");
addressingProperties.setAction(action);
client.getRequestContext().put("javax.xml.ws.addressing.context", addressingProperties);
setClientTimeout(client);
return service;
}
private void setClientTimeout(Client client) {
HTTPConduit conduit = (HTTPConduit) client.getConduit();
HTTPClientPolicy policy = new HTTPClientPolicy();
policy.setConnectionTimeout(applicationProperties.getWex().getServices().getClient().getConnectionTimeout());
policy.setReceiveTimeout(applicationProperties.getWex().getServices().getClient().getReceiveTimeout());
conduit.setClient(policy);
}
This same service bean is accessed by two different threads in the same application sequence. If I execute this particular sequence 10 times in a row, I will get a connection timeout from the service call at least 3 times. What I'm seeing is:
Caused by: java.io.IOException: Timed out waiting for response to operation {http://theservice.com}SearchConsumer.
at org.apache.cxf.endpoint.ClientImpl.waitResponse(ClientImpl.java:685) ~[cxf-core-3.2.0.jar:3.2.0]
at org.apache.cxf.endpoint.ClientImpl.processResult(ClientImpl.java:608) ~[cxf-core-3.2.0.jar:3.2.0]
If I change the sequence such that one of the threads does not call this service, then the error goes away. So, it seems like there's some sort of a race condition happening here. If I look at the logs in our proxy manager for this service, I can see that both of the service calls do return a response very quickly, but the second service call seems to get stuck somewhere in the code and never actually lets go of the connection until the timeout value is reached. I've been trying to track down the cause of this for quite a while, but have been unsuccessful.
I've read some mixed opinions as to whether or not CXF client proxies are thread-safe, but I was under the impression that they were. If this actually not the case, and I should be creating a new client proxy for each invocation, or use a pool of proxies?
Turns out that it is an issue with the proxy not being thread-safe. What I wound up doing was leveraging a solution kind of like one posted at the bottom of this post: Is this JAX-WS client call thread safe? - I created a pool for the proxies and I use that to access proxies from multiple threads in a thread-safe manner. This seems to work out pretty well.
public class JaxWSServiceProxyPool<T> extends GenericObjectPool<T> {
JaxWSServiceProxyPool(Supplier<T> factory, GenericObjectPoolConfig poolConfig) {
super(new BasePooledObjectFactory<T>() {
#Override
public T create() throws Exception {
return factory.get();
}
#Override
public PooledObject<T> wrap(T t) {
return new DefaultPooledObject<>(t);
}
}, poolConfig != null ? poolConfig : new GenericObjectPoolConfig());
}
}
I then created a simple "registry" class to keep references to various pools.
#Component
public class JaxWSServiceProxyPoolRegistry {
private static final Map<Class, JaxWSServiceProxyPool> registry = new HashMap<>();
public synchronized <T> void register(Class<T> serviceTypeClass, Supplier<T> factory, GenericObjectPoolConfig poolConfig) {
Assert.notNull(serviceTypeClass);
Assert.notNull(factory);
if (!registry.containsKey(serviceTypeClass)) {
registry.put(serviceTypeClass, new JaxWSServiceProxyPool<>(factory, poolConfig));
}
}
public <T> void register(Class<T> serviceTypeClass, Supplier<T> factory) {
register(serviceTypeClass, factory, null);
}
#SuppressWarnings("unchecked")
public <T> JaxWSServiceProxyPool<T> getServiceProxyPool(Class<T> serviceTypeClass) {
Assert.notNull(serviceTypeClass);
return registry.get(serviceTypeClass);
}
}
To use it, I did:
JaxWSServiceProxyPoolRegistry jaxWSServiceProxyPoolRegistry = new JaxWSServiceProxyPoolRegistry();
jaxWSServiceProxyPoolRegistry.register(ConsumerSupportService.class,
this::buildConsumerSupportServiceClient,
getConsumerSupportServicePoolConfig());
Where buildConsumerSupportServiceClient uses a JaxWsProxyFactoryBean to build up the client.
To retrieve an instance from the pool I inject my registry class and then do:
JaxWSServiceProxyPool<ConsumerSupportService> consumerSupportServiceJaxWSServiceProxyPool = jaxWSServiceProxyPoolRegistry.getServiceProxyPool(ConsumerSupportService.class);
And then borrow/return the object from/to the pool as necessary.
This seems to work well so far. I've executed some fairly heavy load tests against it and it's held up.
So I've been reading about Spring Message Relay (Spring Messaging stuff) capability with a RabbitMQ broker. What I want to achieve is as follows:
Have a service (1), which acts as a message relay between rabbitmq and a browser. This works fine now. I'm using MessageBrokerRegistry.enableStompBrokerRelay to do that.
Have another service (2) on the back-end, which will send a message to a known queue onto RabbitMQ and have that message routed to a specific user. As a sender, I want to have a control over who the message gets delivered to.
Normally, you'd use SimpMessagingTemplate to do that. Problem is though, that the origin of the message doesn't actually have access to that template, as it's not acting as a relay, it's not using websockets and it doesn't hold mapping of queue names to session ids.
One way I could think of doing it, is writing a simple class on the service 1, which will listen on all queues and forward them using simp template. I fell however this is not an ideal way to do it, and I feel like there might be already a way to do it using Spring.
Can you please advise?
This question got me thinking about the same dilemma I was facing. I have started playing with a custom UserDestinationResolver that arrives at a consistent topic naming scheme that uses just the username and not the session ID used by the default resolver.
That lets me subscribe in JS to "/user/exchange/amq.direct/current-time" but send via a vanilla RabbitMQ application to "/exchange/amqp.direct/users.me.current-time" (to a user named "me").
The latest source code is here and I am "registering" it as a #Bean in an existing #Configuration class that I had.
Here's the custom UserDestinationResolver itself:
public class ConsistentUserDestinationResolver implements UserDestinationResolver {
private static final Pattern USER_DEST_PREFIXING_PATTERN =
Pattern.compile("/user/(?<name>.+?)/(?<routing>.+)/(?<dest>.+?)");
private static final Pattern USER_AUTHENTICATED_PATTERN =
Pattern.compile("/user/(?<routing>.*)/(?<dest>.+?)");
#Override
public UserDestinationResult resolveDestination(Message<?> message) {
SimpMessageHeaderAccessor accessor = MessageHeaderAccessor.getAccessor(message, SimpMessageHeaderAccessor.class);
final String destination = accessor.getDestination();
final String authUser = accessor.getUser() != null ? accessor.getUser().getName() : null;
if (destination != null) {
if (SimpMessageType.SUBSCRIBE.equals(accessor.getMessageType()) ||
SimpMessageType.UNSUBSCRIBE.equals(accessor.getMessageType())) {
if (authUser != null) {
final Matcher authMatcher = USER_AUTHENTICATED_PATTERN.matcher(destination);
if (authMatcher.matches()) {
String result = String.format("/%s/users.%s.%s",
authMatcher.group("routing"), authUser, authMatcher.group("dest"));
UserDestinationResult userDestinationResult =
new UserDestinationResult(destination, Collections.singleton(result), result, authUser);
return userDestinationResult;
}
}
}
else if (accessor.getMessageType().equals(SimpMessageType.MESSAGE)) {
final Matcher prefixMatcher = USER_DEST_PREFIXING_PATTERN.matcher(destination);
if (prefixMatcher.matches()) {
String user = prefixMatcher.group("name");
String result = String.format("/%s/users.%s.%s",
prefixMatcher.group("routing"), user, prefixMatcher.group("dest"));
UserDestinationResult userDestinationResult =
new UserDestinationResult(destination, Collections.singleton(result), result, user);
return userDestinationResult;
}
}
}
return null;
}
}
I've searched quite a bit and been unable to find this: Is there a way that a spring websocket stomp server can disconnect a client based on the sessionId (or really based on anything at all)?
It seems to me that once a client connects to a server there is nothing that allows the server to disconnect the client.
Actually using some workarounds you can achieve what you want.
For that you should do:
Use java configuration (not sure if it is possible with XML config)
Extend your config class from WebSocketMessageBrokerConfigurationSupport and implement WebSocketMessageBrokerConfigurer interface
Create custom sub-protocol websocket handler and extend it from SubProtocolWebSocketHandler class
In your custom sub-protocol websocket handler override afterConnectionEstablished method and you will have access to WebSocketSession :)
I've created sample spring-boot project to show how we can disconnect client session from server side:
https://github.com/isaranchuk/spring-websocket-disconnect
You can also disconnect session by implementing a custom WebSocketHandlerDecorator:
#Configuration
#EnableWebSocketMessageBroker
public class WebSocketConfig<S extends ExpiringSession> extends AbstractSessionWebSocketMessageBrokerConfigurer<S> {
#Override
public void configureWebSocketTransport(final WebSocketTransportRegistration registration) {
registration.addDecoratorFactory(new WebSocketHandlerDecoratorFactory() {
#Override
public WebSocketHandler decorate(final WebSocketHandler handler) {
return new WebSocketHandlerDecorator(handler) {
#Override
public void afterConnectionEstablished(final WebSocketSession session) throws Exception {
session.close(CloseStatus.NOT_ACCEPTABLE);
super.afterConnectionEstablished(session);
}
};
}
});
super.configureWebSocketTransport(registration);
}
#Override
protected void configureStompEndpoints(final StompEndpointRegistry registry) {
registry.addEndpoint("/home")
.setHandshakeHandler(new DefaultHandshakeHandler(
new UndertowRequestUpgradeStrategy() // If you use undertow
// new JettyRequestUpgradeStrategy()
// new TomcatRequestUpgradeStrategy()
))
.withSockJS();
}
}
As far as I know the API doesn't provide what you are looking for, on server-side you can only detect disconnect events. If you want to disconnect a certain client I think you must go for a litte workaround, e.g. this one:
Write a client-side javascript function that is able to trigger a disconnect
As soon as your client is connected to the server, generate a client ID in your javascript and send it to the server. Remember the ID on the client, you'll need it in step (4).
At the time you want the server to disconnect the connection to the specific client (identified by the ID), send a message containing the ID back to the client.
Now your client javascript evaluates the message send from the server and decides to call the disconnect function you wrote in step (1).
Your client disconnects itself.
The workaround is a bit cumbersome but it'll work.
I relied on the idea of #Dániel Kis and implemented the websocket session management with the key point of storing websocket sessions for authenticated users in Singleton-like object.
// WebSocketConfig.java
#Configuration
#EnableWebSocketMessageBroker
public class WebSocketConfig implements WebSocketMessageBrokerConfigurer {
#Override
public void configureWebSocketTransport(WebSocketTransportRegistration registration) {
registration.addDecoratorFactory(new WebSocketHandlerDecoratorFactory() {
#Override
public WebSocketHandler decorate(final WebSocketHandler handler) {
return new WebSocketHandlerDecorator(handler) {
#Override
public void afterConnectionEstablished(final WebSocketSession session) throws Exception {
// We will store current user's session into WebsocketSessionHolder after connection is established
String username = session.getPrincipal().getName();
WebsocketSessionHolder.addSession(username, session);
super.afterConnectionEstablished(session);
}
};
}
});
}
}
Class to store websocket users' sessions WebsocketSessionHolder. I use 'synchronized' blocks for thread safety. Actually this blocks are not expensive operations because each of methods (addSession and closeSessions) are used not so often (On establishing and terminating connection). No need to use ConcurrentHashMap or SynchronizedMap here because we perform bunch of operations with the list in these methods.
// WebsocketSessionHolder.java
public class WebsocketSessionHolder {
static {
sessions = new HashMap<>();
}
// key - username, value - List of user's sessions
private static Map<String, List<WebSocketSession>> sessions;
public static void addSession(String username, WebSocketSession session)
{
synchronized (sessions) {
var userSessions = sessions.get(username);
if (userSessions == null)
userSessions = new ArrayList<WebSocketSession>();
userSessions.add(session);
sessions.put(username, userSessions);
}
}
public static void closeSessions(String username) throws IOException
{
synchronized (sessions) {
var userSessions = sessions.get(username);
if (userSessions != null)
{
for(var session : userSessions) {
// I use POLICY_VIOLATION to indicate reason of disconnecting for a client
session.close(CloseStatus.POLICY_VIOLATION);
}
sessions.remove(username);
}
}
}
}
And the final touch - terminating (disconnecting) specified user websocket sessions ("ADMIN" in the example), say in some Controller
//PageController.java
#Controller
public class PageController {
#GetMapping("/kill-sessions")
public void killSessions() throws Exception {
WebsocketSessionHolder.closeSessions("ADMIN");
}
}
In case of xml configuration you can use <websocket:decorator-factories> in the <websocket:transport> of your <websocket:message-broker>.
Create custom WebSocketHandlerDecorator and WebSocketHandlerDecoratorFactory which implement decorate method.
This may seem brief but I am not certain what the implementation would look like in your case. But, I think there are some circumstances that would warrant this workaround/solution:
Set a timeout on the back-end (say 30 seconds):
This is how you would do it with Spring Boot Websocket (and Tomcat):
#Bean
public ServletServerContainerFactoryBean websocketContainer() {
ServletServerContainerFactoryBean container = new ServletServerContainerFactoryBean();
container.setMaxSessionIdleTimeout(MAX_SESSION_IDLE_TIMEOUT);
return container;
}
If you want to keep the session open - continue to send messages or else actively send ping/pongs. In the case that you want the session to disconnect, stop the ping/pong interaction somewhere suitable in you application.
Of course, if you are wanting to disconnect immediately, this doesn't seem to be an appropriate solution. But if you are simply trying to reduce the number of active connections, ping/pong may be a good fit since it keeps a session open only so long as messages are actively being sent, preventing the session from being closed prematurely.
first you have to introduce a class as your User class by inheritance then use it like this:
if (userObject instanceof User) {
User user = (User) userObject;
if (user.getId().equals(userDTO.getId())) {
for (SessionInformation information : sessionRegistry.getAllSessions(user, true)) {
information.expireNow();
}
}
}
I am connection through SockJS over STOMP to my Spring backend. Everything work fine, the configuration works well for all browsers etc. However, I cannot find a way to send an initial message. The scenario would be as follows:
The client connects to the topic
function connect() {
var socket = new SockJS('http://localhost:8080/myEndpoint');
stompClient = Stomp.over(socket);
stompClient.connect({}, function(frame) {
setConnected(true);
console.log('Connected: ' + frame);
stompClient.subscribe('/topic/notify', function(message){
showMessage(JSON.parse(message.body).content);
});
});
}
and the backend config looks more or less like this:
#Configuration
#EnableWebSocketMessageBroker
public class WebSocketAppConfig extends AbstractWebSocketMessageBrokerConfigurer {
...
#Override
public void registerStompEndpoints(final StompEndpointRegistry registry) {
registry.addEndpoint("/myEndpoint").withSockJS();
}
I want to send to the client an automatic reply from the backend (on the connection event) so that I can already provide him with some dataset (e.g. read sth from the db) without the need for him (the client) to send a GET request (or any other). So to sum up, I just want to send him a message on the topic with the SimMessagingTemplate object just after he connected.
Usually I do it the following way, e.g. in a REST controller, when the template is already autowired:
#Autowired
private SimpMessagingTemplate template;
...
template.convertAndSend(TOPIC, new Message("it works!"));
How to achieve this on connect event?
UPDATE
I have managed to make it work. However, I am still a bit confused with the configuration. I will show here 2 configurations how the initial message can be sent:
1) First solution
JS part
stompClient.subscribe('/app/pending', function(message){
showMessage(JSON.parse(message.body).content);
});
stompClient.subscribe('/topic/incoming', function(message){
showMessage(JSON.parse(message.body).content);
});
Java part
#Controller
public class WebSocketBusController {
#SubscribeMapping("/pending")
Configuration
#Override
public void configureMessageBroker(final MessageBrokerRegistry config) {
config.enableSimpleBroker("/topic");
config.setApplicationDestinationPrefixes("/app");
}
... and other calls
template.convertAndSend("/topic/incoming", outgoingMessage);
2) Second solution
JS part
stompClient.subscribe('/topic/incoming', function(message){
showMessage(JSON.parse(message.body).content);
})
Java part
#Controller
public class WebSocketBusController {
#SubscribeMapping("/topic/incoming")
Configuration
#Override
public void configureMessageBroker(final MessageBrokerRegistry config) {
config.enableSimpleBroker("/topic");
// NO APPLICATION PREFIX HERE
}
... and other calls
template.convertAndSend("/topic/incoming", outgoingMessage);
SUMMARY:
The first case uses two subscriptions - this I wanted to avoid and thought this can be managed with one only.
The second one however has no prefix for application. But at least I can have a single subscription to listen on the provided topic as well as send initial message.
If you just want to send a message to the client upon connection, use an appropriate ApplicationListener:
#Component
public class StompConnectedEvent implements ApplicationListener<SessionConnectedEvent> {
private static final Logger log = Logger.getLogger(StompConnectedEvent.class);
#Autowired
private Controller controller;
#Override
public void onApplicationEvent(SessionConnectedEvent event) {
log.debug("Client connected.");
// you can use a controller to send your msg here
}
}
You can't do that on connect, however the #SubscribeMapping does the stuff in that case.
You just need to mark the service method with that annotation and it returns a result to the subscribe function.
From Spring Reference Manual:
An #SubscribeMapping annotation can also be used to map subscription requests to #Controller methods. It is supported on the method level, but can also be combined with a type level #MessageMapping annotation that expresses shared mappings across all message handling methods within the same controller.
By default the return value from an #SubscribeMapping method is sent as a message directly back to the connected client and does not pass through the broker. This is useful for implementing request-reply message interactions; for example, to fetch application data when the application UI is being initialized. Or alternatively an #SubscribeMapping method can be annotated with #SendTo in which case the resulting message is sent to the "brokerChannel" using the specified target destination.
UPDATE
Referring to this example: https://github.com/revelfire/spring4Test how would that be possible to send anything when the line 24 of the index.html is invoked: stompClient.subscribe('/user/queue/socket/responses' ... from the spring controllers?
Well, look like this:
#SubscribeMapping("/queue/socket/responses")
public List<Employee> list() {
return getEmployees();
}
The Stomp client part remains the same.