Part 1 - http://vunvulearadu.blogspot.ro/2016/03/azure-service-fabric-parent-child.html
Part 2 - http://vunvulearadu.blogspot.ro/2016/03/part-2-azure-service-fabric-parent.html
A few weeks ago I so an interesting question on MSDN forum that I think that is pretty common. In this post I will try to give a possible solution to this problem.
What we need
Basically we need to:
Unique identification of each child service
In our Reliable Dictionary we need to identify unique each service. We could generate a unique ID for each service instance. It might work, but we would need to send this ID to each service instance in the moment when we would call RunAsync method.
Another possible solution is to use the CancellationToken that we already have as a key. The instance of cancellation token is known by parent and child we can use it easily as the key.
Using this approach we can have a mechanism that allow us to have a simple and cheap communication between our services. The main flow would look like this:
This solution can be very useful when we need to migrate a heavy services from a monolithic architecture to a solution hosted on a micro-service system.
The only problem with this solution is related to the type of service. We cannot use multiple types of Reliable Services for this because a Reliable Collection can be shared and accessed only by the same service type.
Tomorrow we will see another solution for this problem without having to use Reliable Collection and the same Reliable Service.
Part 1 - http://vunvulearadu.blogspot.ro/2016/03/azure-service-fabric-parent-child.html
Part 2 - http://vunvulearadu.blogspot.ro/2016/03/part-2-azure-service-fabric-parent.html
Part 2 - http://vunvulearadu.blogspot.ro/2016/03/part-2-azure-service-fabric-parent.html
A few weeks ago I so an interesting question on MSDN forum that I think that is pretty common. In this post I will try to give a possible solution to this problem.
Context:
There are multiple instances of the same Restful Stateful Services that are running in parallel. A new instance is created by a 'parent service that also might specifies the action that needs to be executed.
Problem:
The parent service needs to be able to cancel the instance of our Restful Stateless Service based on external factors or based on the current state of the child.
What we need
Basically we need to:
- Store and map all the service instances that are created by a service
- Store in a specific location a state related to them
- Share their state with the parent service
- Give the ability to the parent to cancel a child service
Sharing the state
To be able to share the state between different instances of the same service we can use Reliable Collections. In our case Reliable Dictionary might be a good solution for our problem. We can store as a key value pair the state of each service instance.
Each instance has the ability to update his state directly. Any concurrency problem like temporary inconsistency are resolved out of the box by Reliable Dictionary.
Cancellation
When we create an instance of a Restful Stateful Service we can use RunAsync method. This method allow us to specify a cancellation token. This token can be used by the child service to see if a cancellation was requests.
protected override async Task RunAsync(
CancellationToken cancellationToken)
{
...
}
Unique identification of each child service
In our Reliable Dictionary we need to identify unique each service. We could generate a unique ID for each service instance. It might work, but we would need to send this ID to each service instance in the moment when we would call RunAsync method.
Another possible solution is to use the CancellationToken that we already have as a key. The instance of cancellation token is known by parent and child we can use it easily as the key.
Using this approach we can have a mechanism that allow us to have a simple and cheap communication between our services. The main flow would look like this:
- [Parent] Create a CancellationToken
- [Parent] Add the CancellationToken to our Mapping State dictionary
- [Parent] Start a new instance of your stateless service and give us parameter the CancellationToken
- [Service Instance] Do his logic
- [Service Instance] Add specific information to the Mapping State dictionary to the item with the same CancellationToken
- [Parent] Detect that a condition is TRUE
- [Parent] Get the CancellationToken of the service instance that needs to be canceled
- [Parent] Send the CancellationToken to the service instance
- [Service Instance] Detect the cancellation request and stop
This solution can be very useful when we need to migrate a heavy services from a monolithic architecture to a solution hosted on a micro-service system.
The only problem with this solution is related to the type of service. We cannot use multiple types of Reliable Services for this because a Reliable Collection can be shared and accessed only by the same service type.
Tomorrow we will see another solution for this problem without having to use Reliable Collection and the same Reliable Service.
Part 1 - http://vunvulearadu.blogspot.ro/2016/03/azure-service-fabric-parent-child.html
Part 2 - http://vunvulearadu.blogspot.ro/2016/03/part-2-azure-service-fabric-parent.html
Comments
Post a Comment