Is it possible to publish a message to only one client subscribed to channel? I know how to get the address (client list). Is it possible to get the address and publish to only that client?
In Redis you publish to a channel, not to a subscriber.
So the only way should be having one channel per client.
You can probably have one common channel to share the address, and then use the addresses as the channels name.
http://redis.io/topics/pubsub
As an example, using C# and StackExchange.Redis library:
//SERVER
var cnn = ConnectionMultiplexer.Connect("localhost");
var sub = cnn.GetSubscriber();
sub.Subscribe("common", (channel, value) =>
{
Console.WriteLine("SERVER Received in common: {0}", value);
sub.Publish("channel:" + value, "private from server to " + value);
});
//CLIENT
var cnn = ConnectionMultiplexer.Connect("localhost");
var address = "client1";
var sub = cnn.GetSubscriber();
sub.Subscribe("channel:" + address, (channel, value) =>
{
Console.WriteLine("Private: {0}", value);
});
sub.Publish("common", address);
Another solution: what I was trying to do was to have all subscribers listen but only one handle the message. The way I did this was create a stack in redis and have all subscribe to a key's list command notifications. So I would RPUSH a new value onto the key, the subscribers would get the keyspace notification and the fist to LPOP the value would handle it.
Related
onUserJoined provides UID which is then used to set up remote video.
in live stream feature when there are two hosts streaming in one channel and if another audience joins later then how can that audience tell which uid is for which host.
Note that onUserJoined does call in random order.
https://docs.agora.io/en/All/API%20Reference/java/classio_1_1agora_1_1rtc_1_1_i_rtc_channel_event_handler.html#a65fd197a39824219aedc2cba81296e82
Use case:
2 hosts already streaming in one channel. 1 of the host is original host and another is a co-host(got invited into the channel)
an audience joins the channel. onUserJoined callback is triggered twice and given the remote UID. That audience now needs to know who is the original host and related his UID and who is the co-host and his related UID.
You basically Answered your self
LIVE_BROADCASTING profile: This callback notifies the app when the host joins the channel. If other hosts are already in the channel, the SDK also reports to the app on the existing hosts. We recommend having at most 17 hosts in a channel.
I finally found a solution for that issue by informing all users that will enter the channel after the Orignal Host by the uid of the orignal host
and you can do this by using sendStreamMessage() function
First joinChannelSuccess triggers after original host login
which will return uid of the original host - and we will save it in a variable.
Then, use sendStreamMessage() to send this uid for all new users
and that by using this function which will be called when userJoined triggers
check this code
int _orignalHost = 'the uid';
_addListener() {
_engine.setEventHandler(RtcEngineEventHandler(
**joinChannelSuccess:** (channel, uid, elapsed) {
setState(() {
isJoined = true;
});
},
**userJoined:** (uid, elapsed) {
remoteUids.add(uid);
setState(() {});
},
// this trriger when you recive the mesage
**streamMessage:** (int uid, int streamId, Uint8List data) {
// save it in _orignalHost variable for the audince
},
streamMessageError:
(int uid, int streamId, ErrorCode error, int missed, int cached) {
},
));
}
Future<void> _onNewUserJoin() async {
try {
var streamId = await _engine
.createDataStreamWithConfig(DataStreamConfig(false, false));
if (streamId != null) {
await _engine.sendStreamMessage(
streamId, Uint8List.fromList(utf8.encode(_orignalHost)));
}
_controller.clear();
} catch (e) {
}
}
you can check full code for StreamMessage:
https://github.com/AgoraIO/Agora-Flutter-SDK/blob/master/example/lib/examples/advanced/stream_message/stream_message.dart
Also, You can check this for more info (anoter solution)
https://docs.agora.io/en/Real-time-Messaging/faq/audience_event?platform=Android
I'm trying to delete a scheduled job in ActiveMQ, and so far no luck.
Schedule is created among with the message using either NMS API or Amqpnetlite (except openwire lib as that one is not updated and can't be used on netstandard/netcore)
Sample code used to create the schedule with NMS, same thing done with AMQP lib:
var factory = new Apache.NMS.ActiveMQ.ConnectionFactory(brokerUri);
IConnection connection = factory.CreateConnection(user, password);
connection.Start();
ISession session = connection.CreateSession(AcknowledgementMode.AutoAcknowledge);
IDestination dest = session.GetQueue(destination);
IMessageProducer producer = session.CreateProducer(dest);
producer.DeliveryMode = MsgDeliveryMode.NonPersistent;
var msg = session.CreateTextMessage("Sample text message");
msg.Properties.SetString("AMQ_SCHEDULED_CRON", "* * * * *");
producer.Send(msg);
connection.Close();
This part produces the following result in browser console, and that's what I am willing to delete:
I've read this other question and answer, also active mq system constants but no way the schedule gets deleted. Also tried to browse over the documentation but can't find anything useful so far
Does ActiveMQ even supports a programmative way of managing schedules? An AMQP solution would be great, but NMS is also appreciated.
You can manage scheduled jobs in ActiveMQ over STOMP, AMQP or simply from a JMS client. I've written about this before showing how to do it using the ActiveMQ Java client but the principle is the same. You can send messages with specific headers set that will operate on the scheduled messages.
To browse the collection of scheduled message you'd do something like the following:
Connection connection = createConnection();
Session session = connection.createSession(false, Session.AUTO_ACKNOWLEDGE);
// Create the Browse Destination and the Reply To location
Destination requestBrowse = session.createTopic(ScheduledMessage.AMQ_SCHEDULER_MANAGEMENT_DESTINATION);
Destination browseDest = session.createTemporaryQueue();
// Create the "Browser"
MessageConsumer browser = session.createConsumer(browseDest);
connection.start();
// Send the browse request
MessageProducer producer = session.createProducer(requestBrowse);
Message request = session.createMessage();
request.setStringProperty(ScheduledMessage.AMQ_SCHEDULER_ACTION,
ScheduledMessage.AMQ_SCHEDULER_ACTION_BROWSE);
request.setJMSReplyTo(browseDest);
producer.send(request);
Message scheduled = browser.receive(5000);
while (scheduled != null) {
// Do something clever...
}
The returned messages contain the information on the actual scheduled message jobs that have been previously added. Obtaining the job Id allows you to completely cancel the delivery of said message.
To remove a scheduled message send that was scheduled using the Java client, AMQP client or other procotol client you'd do the following:
Message remove = session.createMessage();
remove.setStringProperty(ScheduledMessage.AMQ_SCHEDULER_ACTION,
ScheduledMessage.AMQ_SCHEDULER_ACTION_REMOVE);
remove.setStringProperty(ScheduledMessage.AMQ_SCHEDULED_ID,
scheduled.getStringProperty(ScheduledMessage.AMQ_SCHEDULED_ID));
producer.send(remove);
The full set of message property values that can be used when working with the scheduler are documented here, in AMQP just use the string literal of each as the Application Property value you set with the job Id to remote, or in the NMS client it'd just be a string key message property with the job Id that you want to remove.
There is one caveat though when doing this over AMQP and that would be that you'd need to ensure that the broker is using the JMS transformer ?transport.transformer=jms"see the AMQP documentation for ActiveMQ 5.
Accepted answer is working and correct. Turns out scheduledMessage.NMSMessageId was not holding the scheduler id.
Here's the clean C# source in case one is interested:
var factory = new Apache.NMS.ActiveMQ.ConnectionFactory(brokerUri);
IConnection connection = factory.CreateConnection(userName, password);
var session = connection.CreateSession(AcknowledgementMode.AutoAcknowledge);
var requestBrowse = session.GetTopic("ActiveMQ.Scheduler.Management");
var queue = session.GetQueue(queueName);
var consumer = session.CreateConsumer(queue);
connection.Start();
var producer = session.CreateProducer(requestBrowse);
var scheduledMessage = consumer.Receive(TimeSpan.FromSeconds(10));
if (scheduledMessage != null)
{
// do check with persistent storage, if schedule is canceled, remove it:
var remove = session.CreateMessage();
// get prop names from : http://activemq.apache.org/maven/apidocs/constant-values.htm
remove.Properties["AMQ_SCHEDULER_ACTION"] = "REMOVE";
remove.Properties["scheduledJobId"] = scheduledMessage.Properties.GetString("scheduledJobId");
producer.Send(remove);
}
producer.Close();
session.Close();
connection.Close();
Flow is like this: Get messages from some queue, in case some condition is met, drop the schedule entirely.
If I listen event in redis, just like command below,
$ redis-cli --csv psubscribe '__key*__:*'
Reading messages... (press Ctrl-C to quit)
I can get things like
"pmessage","__key*__:*","__keyspace#0__:exipre","xxx"
"pmessage","__key*__:*","__keyevent#0__:expired","xxx"
So, when I use spring-data-redis to listen events in redis, how can I get the specific event name(like keyspace#0:expire), so that not only get the key value like "xxx", but also get the event like expire or expired, as I don't want to build two topic listener respectively like below:
RedisMessageListenerContainer container = new RedisMessageListenerContainer();
container.setConnectionFactory(connectionFactory);
//one for expire
container.addMessageListener(listenerAdapter1, new PatternTopic(//topic 1));
//another for expired
container.addMessageListener(listenerAdapter2, new PatternTopic(//topic 2));
You might want to have a look at KeyspaceEventMessageListener that subscribes to __keyevent#* and then just implement doHandleMessage. Message.getChannel() holds the raw information about the event type like __keyevent#0__:expired.
KeyspaceEventMessageListener eventListener = new KeyspaceEventMessageListener(container) {
#Override
protected void doHandleMessage(Message message) {
String channelInfo = new String(message.getChannel()); // __keyevent#0__:expired
// ...
}
};
I would like to get an exception when sending to a RabbitMQ exchange using MassTransit if the message is not delivered to any queues. It looks like the RabbitMQ mandatory flag could be used for something like this, but I have not been able to find any way to set this using MassTransit. Is this supported?
My code:
var personSendEndpoint = busControl.GetSendEndpoint(new Uri($"rabbitmq://localhost/{personQueueName}")).Result;
// An exception should occur on the following call if the message is not delivered to any queue
personSendEndpoint.Send(new PersonUpdated() { Name = "Dorian Gray", Key = Guid.NewGuid() });
You can set the Mandatory flag, but there isn't a cute extension methods to do it for you. Basically, you do as shown below:
await endpoint.Send(new PersonUpdate(...), context =>
{
(context as RabbitMqSendContext).Mandatory = true;
});
And that will set the flag telling RabbitMQ to throw an exception if no bindings for the exchange exist.
I'm new to RabbitMQ and was wondering of a good approach to this problem I'm mulling over. I want to create a service that subscribes to a queue and only pulls messages that meet a specific criteria; for instance, if a specific subject header is in the message.
I'm still learning about RabbitMQ, and was looking for tips on how to approach this. My questions include: how can the consumer pull only specific messages from the queue? How can the producer set a subject header in the message (if that's even the right term?)
RabbitMQ is perfect for this situation. You have a number of options to do what you want. I suggest reading the documentation to get a better understanding. I would suggest that you use a topic or direct exchange. Topic is more flexible. It goes like this.
Producer code connects to the RabbitMQ Broker and creates and Exchange with a specific name.
Producer publishes to exchange. Each message published will be published with a routing key.
Consumer connects to RabbitMQ broker.
Consumer creates Queue
Consumer binds Queue to the exchange, the same exchange defined in the producer. The binding also includes the routing keys for each message require for this particular consumer.
Lets say you were publishing log messages. The routing key might be something like "log.info", "log.warn", "log.error". Each message published by the producer will have the relevant routing key attached. You will then have a consumer that sends and email for all the error messages and another one that writes all the error messages to a file. So the emailer will define the binding from its queue to the exchange with the routing key "log.error". This way though the exchange receives all messages, the queue defined for the emailer will only contain the error messages. The filelogger will define a new separate queue bound to the same exchange and set up a different routing key. You could do three separate bindings for the three different routing keys require or just use the wildcard "log.*" to request all messages from the exchange starting with log.
This is a simple example that shows how you can achieve what you want to do.
look here for code examples specifically number tutorial number 5.
Making the best of exchange/routing of rabbitmq is recommended. If you do want to check according to the message content, the following code is a viable solution.
Retrieve messages from a queue and check, selectively ack the messages in which you're interested.
pull one message
GetResponse resp = channel.basicGet(QUEUE_NAME, false);
ack one message
channel.basicAck(resp.getEnvelope().getDeliveryTag(), false);
Example
import com.rabbitmq.client.*;
public class ReceiveLogs {
private final static String QUEUE_NAME = "hello";
public static void main(String[] argv) throws Exception {
ConnectionFactory factory = new ConnectionFactory();
factory.setHost("localhost");
try(Connection connection = factory.newConnection();
Channel channel = connection.createChannel();){
channel.queueDeclare(QUEUE_NAME, true, false, false, null);
// pull one message and ack manually and exit
GetResponse resp = channel.basicGet(QUEUE_NAME, false);
if( resp != null ){
String message = new String(resp.getBody(), "UTF-8");
System.out.println(" [x] Received '" + message + "'");
channel.basicAck(resp.getEnvelope().getDeliveryTag(), false);
}
System.out.println();
}
}
}
dependency
compile group: 'com.rabbitmq', name: 'amqp-client', version: '5.8.0'
To Retrieve Message from RabbitMQ we need to first connect with RabbitMQ server
public WebClient GetRabbitMqConnection(string userName, string password)
{
var client = new WebClient();
client.Credentials = new NetworkCredential(userName, password);
return client;
}
Now retrieve message from RabbitMQ using below code.
public string GetRabbitMQMessages(string domainName, string port,
string queueName, string virtualHost, WebClient client, string methodType)
{
string messageResult = string.Empty;
string strUri = "http://" + domainName + ":" + port +
"/api/queues/" + virtualHost + "/";
var data = client.DownloadString(strUri + queueName + "/");
var queueInfo = JsonConvert.DeserializeObject<QueueInfo>(data);
if (queueInfo == null || queueInfo.messages == 0)
return string.Empty;
if (methodType == "POST")
{
string postbody = "
{\"ackmode\":\"ack_requeue_true\",\"count\":
\"$totalMessageCount\",\"name\":\"${DomainName}\",
\"requeue\":\"false\",\"encoding\":\"auto\",\"vhost\" :
\"${QueueName}\"}";
postbody = postbody
.Replace("$totalMessageCount", queueInfo.messages.ToString())
.Replace("${DomainName}", domainName)
.Replace("${QueueName}", queueName);
messageResult = client.UploadString(strUri + queueName +
"/get", "POST", postbody);
}
return messageResult;
}
I think this will help you to implement RabbitMQ.
If you want to retrieve single message at a time please add the following properties with your Retrieving code .
Boolean autoAck = false;
model.BasicConsume(Queuename, autoAck);
model.BasicGet("Queuename", false);
model.BasicGet("Queuename", false);
By adding this properties of RabbitMQ you can retrieve the message one by one from the queue .Same like FIFO criteria