I have regular net.tcp WCF service client, and regular net.tcp duplex (i.e. with a callback) WCF service client. I have implemented some logic to constantly reinstantiate the connection in case a service has faulted.
They are created in the exactly same way:
FooServiceClient Create()
{
var client = new FooServiceClient(ChannelBinding);
client.Faulted += this.FaultedHandler;
client.Ping(); // An empty service function to make sure connection is OK
return client;
}
BarServiceClient Create()
{
var duplexClient = new BarServiceClient(new InstanceContext(this.barServiceCallback));
duplexClient.Faulted += this.FaultedHandler;
duplexClient.Ping(); // An empty service function to make sure connection is OK
return duplexClient;
}
public class Watcher
{
public Watcher()
{
this.CommunicationObject = this.Create();
}
ICommunicationObject CommunicationObject { get; private set; }
void FaultedHandler(object sender, EventArgs ea)
{
this.CommunicationObject.Abort();
this.CommunicationObject.Faulted -= this.FaultedHandler;
this.CommunicationObject = this.Create();
}
}
The FaultedHandler()
aborts the channel and recreates it using the code above.
The FooServiceClient
reconnection logic works just fine, it is being reconnected after many faults. Whereas, almost the same but duplex BarServiceClient
receives Faulted event only from the first BarServiceClient
instance, i.e once.
Why only the firsts instance of duplex BarServiceClient
gets faulted event? Are there any workarounds?
A similar non answered question: WCF Reliable session without transport security will not faulted event on time