Problem Statement

Currently, when a developer wants to change the configuration of a cluster, say, create a region (or destroy an index, or update an async event queue) and wants to have the change persisted in the cluster configuration for incoming servers, there is no public API for them to do so. They must replicate the effort of the equivalent gfsh command to achieve the same effect. It would be nice if we can expose what these commands do to a public API.

Product Goals 

User Goals

Create a more modular product to allow for easy extension and integration.

The beneficiaries of this work are those who want to change the configuration of the cluster (create/destroy regions, indices or gateway receivers/senders etc), and have these changes replicated on all the applicable servers and persisted in the configuration persistence service for new joining servers. This includes developers working on different parts of the code such as Spring Data for Apache, Queries for Lucene index, storage for the JDBC connector, and other Geode developers.

What We Have Now:

Our admin rest API "sort of" already serves this purpose, but it has these shortcomings:

1. It's not a public API
2. The API is restricted to the operations implemented as gfsh commands, as the argument to the API is a gfsh command string.
3. Each command does similar things, yet commands may not be consistent with each other.

Below is a diagram of the current state of things:

From the current state of commands, It's not easy to extract a common interface for all the commands. And developers do not want to use gfsh command strings as a "makeshift" API to call into the command. We are in need of a unified interface and a unified workflow for all the commands.

Proposal

We propose a new Cluster Management Service (CMS) which has two responsibilities:

• Update runtime configuration of servers (if any running)
• Persist configuration (if enabled)

Note that in order to use this API, Cluster Configuration needs to be enabled.

The CMS API is exposed as a new endpoint as part of "Admin REST APIs", accepting configuration objects (JSON) that need to be applied to the cluster. CMS adheres to the standard REST semantics, so users can use POST, PATCH, DELETE and GET to create, update, delete or read, respectively. The API returns a JSON body that contains a message describing the result along with standard HTTP status codes.

Root End Point

{
    "number_of_locators": 3,
	"number_of_servers": 8,
	"region_url": "/geode/v2/regions",
	"gateway_receiver_url": "/geode/v2/gwr",
	"gateway_sender_url": "/geode/v2/gws"
}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for CLUSTER:READ"
}

Create End Point

{
  "regionConfig": {
      "name": "Foo",
      "type": "REPLICATE" 
  }
}

{
  "Metadata": {
    "Url": "/geode/v2/regions/Foo"
  }
}

{
  "message": "Region /Foo already exists"
}

{
    "message": "Region type is a required parameter"
}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for DATA:MANAGE"
}

{
    "message": "Failed to create region /Foo because of <reason>"
}

Note that the CREATE endpoint is idempotent – i.e. it should be a NOOP if the region already exists.

List End Point

{
    "Total_results": 10,
    "Regions" : [
     {
       "Name": "Foo",
       "Url": "/geode/v2/regions/Foo"
     },
     ...
     ]
}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for CLUSTER:READ"
}

Describe End Point

{
    "Name": "Foo",
    "Data_Policy": "partition",
    "Hosting_Members": [
      "s1",
      "s2",
      "s3"
      ],
    "Size": 0,
    "Indices": [
     {
     "Id": 111,
     "Url": "/geode/v2/regions/Customer/index/111"
     }
    ]

}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for CLUSTER:READ"
}

{
     "message": "Region with name '/Foo' does not exist"
}

Update End Point

{
  "regionConfig": {
      "gateway_sender_id": ["1","2"]
  }
}

{
  "Metadata": {
    "Url": "/geode/v2/regions/Foo"
  }
}

{
    "message": "Invalid parameter specified"
}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for DATA:MANAGE"
}

{
    "message": "Region with name '/Foo' does not exist"
}

{
    "message": "Failed to update region /Foo because of <reason>"
}

Delete End Point

{
    "message": "Region with name '/Foo' does not exist"
}

{
    "message": "Missing authentication credential header(s)"
}

{
    "message": "User1 not authorized for DATA:MANAGE"
}

{
    "message": "Failed to delete region /Foo because of <reason>"
}

Note that the DELETE endpoint is idempotent – i.e. it should be a NOOP if the region does not exist.

Let's look at some code to see how users can use this service. The below example shows how to create a region using CMS.

Curl (any standard REST client)

curl http://locator.host:8080/geode/v2/regions -XPOST -d '
{
  "regionConfig": {
      "name": "Foo" 
      "type": "PARTITIONED"
  }
}'

On Client 

public class MyApp {
  public static void main(String[] args) {
    //1. Get the service from Cache
    ClientCache cache = new ClientCacheFactory().addPoolLocator("127.0.0.1", 10334).create();
    ClusterManagementService cms = cache.getClusterManagementService();
    
    //2. Create the config object, these are just JAXB generated POJOs
    RegionConfig regionConfig = new RegionConfig(); //These are JAXB generated configuration objects
    regionConfig.setrefId("REPLICATE");
    
    //3. Invoke create, update, delete or get depending on what you want to do.
    ConfigResult result = cms.createRegion("Foo", regionConfig); //create(regionName, config) returns a ConfigResult or throws an exception   
  }
}

On Server

Here's how one can use CMS on a server.

public class MyFunction implements Function<String> {
  @Override
  public void execute(FunctionContext context) {
    //1. Get the service from cache
 	Cache cache = context.getCache();
    ClusterConfigurationService cms = Cache.getClusterManagementService();
    
    //2. Create the config object, these are just JAXB generated POJOs
    RegionConfig regionConfig = new RegionConfig(); //These are JAXB generated configuration objects
    regionConfig.setrefId("REPLICATE");
    
    //3. Invoke create, update, delete or get depending on what you want to do.
    ConfigResult result = cms.createRegion("Foo", regionConfig); //create(regionName, config) returns a ConfigResult or throws an exception
  }
}

Behind the scenes

Following the effort here, Configuration Persistence Service, we already have a set of configuration objects derived from the cache XML schema. This would serve a common object that the developer would use to configure the config instance. The developer would then ask the cluster management service to persist it, either on the cache (creating the real thing on an existing cache) or on the configuration persistence service (persisting the configuration itself). 

Pros and Cons:

Pros:

1. A common interface to call either on the locator/server/client side
2. A common workflow to enforce behavior consistency
3. Modularized implementation. The configuration object needs to implement the additional interfaces in order to be used in this API. This allows us to add functionality gradually and per function groups.

Cons:

1. Existing gfsh commands need to be refactored to use this API as well, otherwise we would have duplicate implementations, or have different behaviors between this API and gfsh commands.
2. When refactoring gfsh commands, some commands' behaviors will change if they want to strictly follow this workflow, unless we add additional APIs for specific configuration objects.

Migration Strategy:

Our current commands uses numerous options to configure the behavior of the commands. We will have to follow these steps to refactor the commands.

1. Combine all the command options into one configuration object inside the command itself.
   
2. Have the command execution call the public API if the command conforms to the new workflow. In this step, the config objects needs to implement the ClusterConfigElement.
3. If the command can't use the common workflow, make a special method in the API for that specific configuration object. (We need to evaluate carefully - we don't want to make too many exceptions to the common workflow.)

The above work can be divided into functional groups so that different groups can share the workload.

Once all the commands are converted using the ClusterManagementService API, each command class can be reduced to a facade that collects the options and their values, builds the config object and calls into the API. At this point, the command objects can exist only on the gfsh client.

The end architecture would look like this:

Project Milestones

1. API is clearly defined
2. All commands are converted using this API
3. Command classes exist only on a gfsh client. The GfshHttpInvoker uses the REST API to call this ClusterConfigurationService with the configuration objects directly.