Tutorials for nyssr.net on sillysky.net

How to build a message

Tue, 09 Dec 2025 00:00:00 +0000

A message consists of an envelope and a record. The envelope contains data that controls the transport of the message to its recipient, similar to a letter. The record is the payload of the message. It consists of a map of keys and values. A key-value pair is called a slot. In addition, the record has an ID, also known as a message ID.

A cumbersome but working way of creating messages is to assemble them manually. The problem remains that the recipient must know exactly how the message is structured in order to access the content.

How to register a nano service

Tue, 09 Dec 2025 00:00:00 +0000

Registration of the nano service

Before you can use nano services, they should be registered. This can be done manually by the operator of the service, or automatically in a ServiceStarter.

You should have read these tutorials beforehand:
Package initialization
How to start a simple target

Registration manually

Nano services are always bound to a namespace. In most cases, this is the SYSTEM namespace. So we need the nano service registry of the namespace. Preferably we register the nano service in the StartTarget handler (see How to start a simple target).

How to use the job engine for bulk tasks

Tue, 09 Dec 2025 00:00:00 +0000

The job machine is designed to run many jobs in parallel as fast as possible. In this tutorial we want to calculate the MD hash values for each file in a directory with many files.

The example is a bit longer because many tasks are presented here:

The definition of dependencies
The service starter
The operation of a nano service
The outsourcing of the request to a working target, in order that parallel requests can be processed
The initialization of the job machine
The job itself
The orderly termination of the machine in case of success and in case of termination

The service target

To make the exercise more understandable, we register a nanoservice. Creating hashes of the files in a directory is thus possible by sending a message to this nanoservice.

How to send a message

Fri, 07 Nov 2025 00:00:00 +0000

Sending a message to a known receiver

A message can be sent directly to a receiver if its target address is known. It is stored in the envelope, as is the sender's address.

How do I edit the settings in the Java Preferences?

Wed, 15 Oct 2025 00:00:00 +0000

The use of Java Preferences does not seem to be very common anymore. Thereby this is a very useful library, which is also included in the runtime with Java. Unfortunately, the tools around this good framework are correspondingly old.

An editor intended for non-commercial use only (GPL) is the javaprefs editor (prefs-0.8.jar) by Boris Gontar. I have linked the file for download here, since I can no longer find it on the Internet.

How new software is installed on a node remotely

Wed, 15 Oct 2025 00:00:00 +0000

The software of a node can be updated semi-automatically.

A node consists of a kernel and the plugins that this kernel loads. Additionally, there is third-party software in the form of libraries. All components are located in two directories. The lib directory contains the kernel, the API JARs of the plugins, and the third-party software. The classpath contains only one entry, namely the lib directory. The plugins directory contains the plugin JAR files.

How nodes are connected to each other

Wed, 15 Oct 2025 00:00:00 +0000

Nodes are connected to each other through TCP channels. Through the channels, the nodes communicate with each other. These connections are open all the time.

To create the TCP channels, nodes must act as TCP servers or TCP clients. Many nodes have both functions as well. For a node to start the server or client, an entry in the configuration is sufficient.

Here is a sample configuration file for a TCP server.

How to add an observer to a nano service

Wed, 15 Oct 2025 00:00:00 +0000

Registering observers to nano services is quite common.

Prerequisite for the tutorial

You should have read these tutorials beforehand:
Package initialization
How to start a simple target

The short way

To make registration as easy as possible, another service has been created: the RecordHelper. Enter the dependency in the ServiceStarter.

public void getDependencies(@NotNull final IServiceDependencyList aDependencyList)
{
    aDependencyList.add(IRecordHelper.class);
}

In the target:

private boolean asyncStartTarget(@NotNull final CEnvelope aEnvelope,
                                 @NotNull final CRecord aRecord) throws CException
{
    mDependencies.getRecordHelper()
                 .addObserver(CRecordNotifyRemoteNodeAdded.class,
                              this,
                              true);
    aEnvelope.setResultSuccess();
    return true;
}

The long way

Add the nano service registry as a dependency to the service starter. As a filter, we need to enter the namespace where the desired nano service is registered.

How to catch a message

Wed, 15 Oct 2025 00:00:00 +0000

A message sent to a target can be handled by the target in a message handler. The message handler must be registered for this—preferably in the constructor of the target.

CMyTarget()
{
    addMessageHandler(CRecordMyMessage.ID,
                      this::asyncMyMessage);
}

Here the message CRecordMyMessage is handled by the handler asyncMyMessage.

private boolean asyncMyMessage(@NotNull final CEnvelope aEnvelope,
                               @NotNull final CRecord aRecord)
{
    if (aEnvelope.isAnswer())
    {
        // do something...

        return false;
    }
    else
    {
        // do something...

        aEnvelope.setResultSuccess();
        return true;
    }
}

Requests are processed as well as responses in a single handler. It is safer to always check the message for the response flag, even if a request or response is not expected.

How to create a microservice

Wed, 15 Oct 2025 00:00:00 +0000

Microservices in nyssr.net are usually smaller than the variants in the restful ecosystem. The reason for this is that very many microservices can be run by a single node.

Microservices can also communicate in all directions, a consequence of the way in which messaging is used. They can also be factories to create instances of services that communicate exclusively with an owner.

Therefore, a microservice always consists of at least one target so that it can communicate.

How to create an application

Wed, 15 Oct 2025 00:00:00 +0000

In the nyssr.net context, there are usually many more applications than in the normal server environment. This is because the applications and microservices are much smaller than in the Restful environment, due to the fact that a single server (called node in our context) can run hundreds or thousands of modules. To us, applications are always factory microservices that create and launch the application instances.

How to import files into nyssr.net

Wed, 15 Oct 2025 00:00:00 +0000

To work with files in nyssr.net, two plugins are required.

The NY_FileRegistryPlugIn plugin runs as a singleton on a central node, meaning it is loaded on only one node. It creates the The File Registry API microservice, which is used internally to locate files within nyssr.net.

The NY_FileStorePlugIn plugin is loaded on every node that provides or requires files. It manages files in its working directory, which is usually located in the storage directory (storage/FileStore/indexed/). The files are stored in a decentralized manner, so different or identical files may be available in the various FileStores.

How to offer a nano service

Wed, 15 Oct 2025 00:00:00 +0000

Nano services are needed for three purposes:

Offering notifications.

For a nano service, interested targets can register an observer. The notification trigger sends messages to the nano service, of which each observer receives a copy.

Operation of a tiny services

The owner registers a nano service and registers himself as an observer. Clients of the service send a message to the nano service, which is forwarded to the owner. Advantage: The client does not need the address of the providing service.

How to start a node

Wed, 15 Oct 2025 00:00:00 +0000

A node requires some information without which it cannot be executed. Besides the classpath for the kernel JAR file and the libraries this includes:

node.id (the node ID)
config.dir (directory for configuration files)
plugin.dir (directory for plugins)
storage.dir (directory for storage)
segment.id (the segment ID)
type.of.node (the type of the node)
license (the license code)

All data except the node ID can be taken from the Java preferences, provided that the node has already been executed correctly. However, for a very first launch, we need to provide this data to the Node in a different way. One way is to use configuration files.

How to start a simple target

Wed, 15 Oct 2025 00:00:00 +0000

Short variant

A simple target is a class derived from CTarget. This target must be registered to the target registry of a namespace. There it receives its target address and can receive messages from now on.

final INamespace namespace = mDependencies.getNamespaceRegistry()
                                          .getNamespace(CWellKnownNID.SYSTEM);
if (namespace != null)
{
    namespace.getTargetRegistry()
             .registerTarget(this);
}

In order to use the mDependencies interface, as always, a few preparations must be made. For this we refer to the Tutorial on initializing a Java package.

Package initialization

Wed, 15 Oct 2025 00:00:00 +0000

Packages are used in Java to group all classes belonging to a module in one place. Only the interface to the module should be public, all other classes and methods remain private.

In many cases, a module has dependencies on other modules or services. Now you could check before each use if all dependencies are fulfilled, i.e. the corresponding modules are available. However, this is a laborious approach.

The use of plugins

Wed, 15 Oct 2025 00:00:00 +0000

We decided to outsource services to plugins. We can of course load normal JARS at startup, but plugins have some advantages. So the application of the node is always the same - we load all node-specific functions via plugins. This comes into play in containers at the latest when plugins are loaded via a shared volume. One image, many nodes, and the behavior of the nodes is controlled by a list in the configuration. It couldn't be simpler. Well, we still have a few ideas, wait and see!

The use of timer messages

Wed, 15 Oct 2025 00:00:00 +0000

Timers in nyssr.net are messages that are sent to their target after a certain time. They can be sent repeatedly, and they can transport user data.

As always, we assume that the package has been initialized. Here are the dependencies we have:

@Override
public void getDependencies(@NotNull final IServiceDependencyList aDependencyList)
{
    aDependencyList.add(INamespaceRegistry.class);
    aDependencyList.add(ITimerManager.class);
}

We need the namespace registry to fetch a namespace in which we can register the target. For the timer we need the Timer Manager.

Tutorials for nyssr.net on sillysky.net

How to build a message

How to register a nano service

Registration of the nano service

Registration manually

How to use the job engine for bulk tasks

The service target

How to send a message

Sending a message to a known receiver

How do I edit the settings in the Java Preferences?

How new software is installed on a node remotely

How nodes are connected to each other

How to add an observer to a nano service

Prerequisite for the tutorial

The short way

The long way

How to catch a message

How to create a microservice

How to create an application

How to import files into nyssr.net

How to offer a nano service

Offering notifications.

Operation of a tiny services

How to start a node

How to start a simple target

Short variant

Package initialization

The use of plugins

The use of timer messages