Resiliency in Contrail Service Orchestration
The Contrail Services Orchestration (CSO) software offers robust deployment implementations with resiliency for the following features:
High availability of CSO infrastructure services and microservices in medium and large installations
Each infrastructure service or microservice resides on multiple hosts and if an application on the primary host fails, a corresponding application on another host takes over. Current operations for an application do not recover if a failure occurs; however, any new operations proceed as normal.
A Contrail OpenStack instance configured for high availability can be used with the centralized Cloud CPE deployment
The Contrail OpenStack instance, which is used for the centralized deployment, includes three Contrail controller nodes in the Contrail Cloud Platform, and provides resiliency for virtualized infrastructure managers (VIMs), virtualized network functions (VNFs), and network services.
CSO provides additional resiliency for virtualized network functions (VNFs) and network services in the centralized Cloud CPE solution.
You can enable or disable automatic recovery of a network service in a centralized deployment. If a network service becomes unavailable due to a connectivity issue with a VNF, Network Service Orchestrator maintains existing instances of the network service in end users’ networks and initiates recreation of the VNFs. During this recovery process, the end user cannot activate the network service on additional network links. When the problem is resolved, normal operation resumes and end users can activate the network service on additional network links.
Enabling automatic recovery improves reliability of the implementation. Conversely, disabling automatic recovery for a network service allows you to quickly investigate a problem with the underlying VNF. By default, automatic recovery of a network service is enabled.
