MassTransit RPC (RabbitMQ) timeout up with multiple clients
I have two node types. A Server and a Client . There will be 1 or more Server or Client currently running.
Using MassTransit and RabbitMQ, I am getting a timeout when trying the RPC pattern described here.
The code for Server and Client is simple, for testing.
Code for Server
class Program
{
static void Main(string[] args)
{
var serviceBus = ServiceBusFactory.New(sbc =>
{
sbc.UseRabbitMq();
sbc.ReceiveFrom("rabbitmq://192.168.10.201/mybus_responder");
sbc.Subscribe(x =>
{
x.Handler<TestEvent>((ctx, @event) =>
{
Console.WriteLine("Received message " + @event.Message);
ctx.Respond(new TestEventResponse { Response = "Processed message "" + @event.Message + """ });
});
x.Handler<TestEvent2>((ctx, @event) =>
{
Console.WriteLine("Received message " + @event.Message);
ctx.Respond(new TestEventResponse2 { Response = "Processed message "" + @event.Message + """ });
});
});
});
Console.WriteLine("Listening for events...");
Console.ReadLine();
}
}
Code for Client
class Program
{
static void Main(string[] args)
{
var serviceBus = ServiceBusFactory.New(sbc =>
{
sbc.UseRabbitMq();
sbc.ReceiveFrom("rabbitmq://192.168.10.201/mybus");
});
Console.WriteLine("Ready to send events...");
var eventType = false;
while (true)
{
eventType = !eventType;
var message = Path.GetFileNameWithoutExtension(Path.GetRandomFileName());
Console.WriteLine("Sending message " + message);
if (eventType)
{
serviceBus.PublishRequest(new TestEvent {Message = message}, configurator =>
{
configurator.Handle<TestEventResponse>(response =>
{
Console.WriteLine(response.Response);
});
configurator.SetTimeout(TimeSpan.FromSeconds(5));
});
}
else
{
serviceBus.PublishRequest(new TestEvent2 {Message = message}, configurator =>
{
configurator.Handle<TestEventResponse2>(response =>
{
Console.WriteLine(response.Response);
});
configurator.SetTimeout(TimeSpan.FromSeconds(5));
});
}
}
}
}
If I run exactly 1 of each node type, the commands/responses perform as expected. If I have more than 1 Server instances, the commmands/responses perform as expected, distributing Client requests evenly between each ```Server`` node.
However, if I have multiple Client instances running, I will get timeouts on the ```Client`` side.
Why is this? I require support for multiple ```Client`` instances, because this is my web-tier.
Do, I believe this is simply because you end up sharing a queue between instances and they gobble each other's messages. Each client and sever needs a unique queue to read from.
链接地址: http://www.djcxy.com/p/61226.html