- play_arrow Overview
- play_arrow Understanding Contrail
-
- play_arrow Installing and Upgrading Contrail
- play_arrow Server Requirements and Supported Platforms
- play_arrow Installing Contrail and Provisioning Roles
- Introduction to Containerized Contrail Modules
- Introduction to Contrail Microservices Architecture
- Downloading Installation Software
- Overview of contrail-ansible-deployer used in Contrail Command for Installing Contrail with Microservices Architecture
- Installing Contrail with OpenStack and Kolla Ansible
- Configuring the Control Node with BGP
- Contrail Global Controller
- Role and Resource-Based Access Control
- play_arrow Installation and Configuration Scenarios
- play_arrow Upgrading Contrail Software
- play_arrow Backup and Restore Contrail Software
- play_arrow Contrail Command
- play_arrow Multicloud Contrail
- play_arrow Using Contrail with Red Hat
- play_arrow Using Contrail with Juju Charms
-
- play_arrow Configuring Contrail
- play_arrow Configuring Virtual Networks
- Creating Projects in OpenStack for Configuring Tenants in Contrail
- Creating a Virtual Network with Juniper Networks Contrail
- Creating a Virtual Network with OpenStack Contrail
- Creating an Image for a Project in OpenStack Contrail
- Creating a Floating IP Address Pool
- Using Security Groups with Virtual Machines (Instances)
- Support for IPv6 Networks in Contrail
- Configuring EVPN and VXLAN
- play_arrow Example of Deploying a Multi-Tier Web Application Using Contrail
- play_arrow Configuring Services
- play_arrow Configuring Service Chaining
- play_arrow Examples: Configuring Service Chaining
-
High Availability for Analytics
Contrail supports multiple instances of analytics for high availability and load balancing.
Contrail analytics provides two broad areas of functionality:
contrail-collector —Receives status, logs, and flow information from all Contrail processing elements (for example, generators) and records them.
Every generator is connected to one of the contrail-collector instances at any given time. If an instance fails (or is shut down), all the generators that are connected to it are automatically moved to another functioning instance, typically in a few seconds or less. Some messages may be lost during this movement. UVEs are resilient to message loss, so the state shown in a UVE is kept consistent to the state in the generator.
contrail-opserver —Provides an external API to report UVEs and to query logs and flows.
Each analytics component exposes a northbound REST API represented by the contrail-opserver service (port 8081) so that the failure of one analytics component or one contrail-opserver service should not impact the operation of other instances.
These are the ways to manage connectivity to the contrail-opserver endpoints:
Periodically poll the contrail-opserver service on a set of analytics nodes to determine the list of functioning endpoints, then make API requests from one or more of the functioning endpoints.
The Contrail user interface makes use of the same northbound REST API to present dashboards, and reacts to any contrail-opserver high availability event automatically.
