Migration Notes

Retired Content

This content and the technology described is outdated and is no longer being maintained. For more information, see Transient Fault Handling.

patterns & practices Developer Center

For users of the Transient Fault Handling Application Framework from the Microsoft Azure Customer Advisory Team, this topic summarizes the key differences that you should be aware of if you are planning to migrate to the Transient Fault Handling Application Block.

The following is a list of the key changes.

  • The RetryPolicy.RetryOccurred event is renamed as RetryPolicy.Retrying. It is refactored to a .NET event with an EventArgs parameter following the .NET design guidelines.
  • The overload of the RetryPolicy.ExecuteAction method that supports asynchronous calls now includes a success delegate that you can use to act on the result of the asynchronous operation.
  • The overload of the RetryPolicy.ExecuteAction method that supports asynchronous calls now detects transient failures that might occur during the Begin operation.
  • The Transient Fault Handling Application Block introduces the concept of a retry strategy that defines how a transient failure should be retried and how long to wait before retrying. The RetryStrategy class follows the extensible provider pattern, enabling you to create custom retry strategies.
  • You can now retrieve retry strategies and retry policies from the RetryManager class. The RetryManager class is resolved from the Enterprise Library container (instead of through the configuration objects). For more information about creating and referencing Enterprise Library objects, see the topic "Creating and Referencing Enterprise Library Objects" on MSDN.
  • The "communication" retry policy has been renamed "Windows Azure Service Bus" retry policy.
  • The non-generic RetryPolicy class is no longer an abstract class. You can now use it directly by providing an instance of an error detection strategy.
  • The classes in the Instrumentation namespace were removed.
  • It is now required that you provide a default retry strategy in the configuration.
  • RetryPolicyFactory.GetRetryPolicy<T>(string retryStrategyName) now throws an ArgumentOutOfRangeException exception if the retry strategy cannot be found in the configuration.
  • There are major changes to the configuration schema. See the topic "Entering Configuration Information" for more details.
  • Many of the default values for configuration entries have changed. See the topic "Entering Configuration Information" for more details.

Next Topic | Previous Topic | Home

Last built: June 7, 2012