Connection refused after upgrade Worker.Extensions.DurableTask 1.4.0 -> 1.5.0

[eholman]

I assume that this most likely isn't a bug but more a configuration issue but I can't find the solution.

When I was upgrading my isolated Functions project with ASP.NET Core Integration that depends on Durable Functions I couldn't schedule a new orchestration instance anymore.

This is the exception I got after upgrading Microsoft.Azure.Functions.Worker.Extensions.DurableTask:

Grpc.Core.RpcException: Status(StatusCode="Unavailable", Detail="Error connecting to subchannel.", DebugException="System.Net.Sockets.SocketException: Connection refused")
 ---> System.Net.Sockets.SocketException (111): Connection refused
   at System.Net.Sockets.Socket.AwaitableSocketAsyncEventArgs.ThrowException(SocketError error, CancellationToken cancellationToken)
   at System.Net.Sockets.Socket.AwaitableSocketAsyncEventArgs.System.Threading.Tasks.Sources.IValueTaskSource.GetResult(Int16 token)
   at Grpc.Net.Client.Balancer.Internal.SocketConnectivitySubchannelTransport.TryConnectAsync(ConnectContext context, Int32 attempt)
   --- End of inner exception stack trace ---
   at Grpc.Net.Client.Balancer.Internal.ConnectionManager.PickAsync(PickContext context, Boolean waitForReady, CancellationToken cancellationToken)
   at Grpc.Net.Client.Balancer.Internal.BalancerHttpHandler.SendAsync(HttpRequestMessage request, CancellationToken cancellationToken)
   at Grpc.Net.Client.Internal.GrpcCall`2.RunCall(HttpRequestMessage request, Nullable`1 timeout)
   at Grpc.Net.Client.Internal.GrpcCall`2.GetResponseHeadersCoreAsync()
   at Microsoft.DurableTask.Client.Grpc.GrpcDurableTaskClient.ScheduleNewOrchestrationInstanceAsync(TaskName orchestratorName, Object input, StartOrchestrationOptions options, CancellationToken cancellation)
   // own lib callstack

So I decided to find out since which version this behavioral change was and my conclusion was 1.5.0. I checked out the compare log between 1.4.0 and 1.5.0 and found this PR but I'm not sure this is the root cause:

#3126

The logs are changed accordingly as per that PR though:

Opened local gRPC endpoint: http://127.0.0.1:38831 (Mode=AspNetCore). InstanceId: . Function: . HubName: TestHubName. AppName: . SlotName: . ExtensionVersion: 3.3.0. SequenceNumber: 0.

This piece of code was working before and broken after the upgrade:

        services.AddDurableTaskClient(builder =>
        {
            builder.UseGrpc();
            builder.RegisterDirectly();
        });

This is my host.json extensions part which didn't change as well

    "extensions": {
        "durableTask": {
            "tracing": {
                "distributedTracingEnabled": true,
                "version": "V2"
            },
            "storageProvider": {
                "type": "mssql",
                "connectionStringName": "EnterpriseResourcePlanningServicesDbConnection",
                "taskEventLockTimeout": "00:02:00",
                "createDatabaseIfNotExists": true,
                "schemaName": "dt"
            }
        }
    },

My durable package dependencies and this are the only packages upgraded:

• .NET 10 (no issues with this upgrade only; coming from .NET 8)
• Microsoft.Azure.Functions.Worker.Extensions.DurableTask.SqlServer 1.5.2 (no issues with this upgrade only; coming from 1.5.0)
• Microsoft.Azure.Functions.Worker.Extensions.DurableTask 1.5.0 (no issues with 1.4.0)

I can't figure out what I must do in order to get it working again but I suspect it has something to do with the UseGrpc() that can't figure out where the 'new' gRPC server is at...