High availability deployment for TrueSight Orchestration in containers
High availability is a redundancy operation that automatically switches to a standby server if the primary server fails or is temporarily shut down for maintenance.
When you choose to install TrueSight Orchestration in containers, you use Docker Swarm for creating host clusters. Using a swarm ensures that your platform environment is in a cluster, supports high availability, and is fault-tolerant. For the configuration distribution peer, you no longer require to install the secondary or High Availability CDP (HA-CDP) component. Swarm consists of hosts or nodes that are designated as manager and worker nodes. Adding a manager node in a swarm makes it more fault-tolerant.
In a swarm mode, failover is handled in the following ways:
- Service stops running: If a component service is down on one of the hosts, swarm ensures that the service is restarted on the same host/node automatically.
- Host is down: If the host is down, swarm restarts the component service on any of the other nodes in the swarm.
BMC recommends that you install TrueSight Orchestration in a minimum three-node swarm cluster with each component installed on a separate node. If, for some reason, the CDP service service stops running, the swarm restarts the service within seconds on the host where the CDP image is installed. If the host or node is down, the CDP service is started on one of the other nodes in a swarm. In addition, you use Network File System (NFS) with swarm for data replication across nodes delegated to the NFS clients.
This section helps you in locating information about planning, installing, and configuring high availability for TrueSight Orchestration in containers.
The following figure depicts the HA environment for TrueSight Orchestration in containers achieved by using Network File System (NFS) with Docker.
|Docker Engine||Docker Engine is the underlying technology that builds and runs containers using Docker components and servers. TrueSight Orchestration uses Docker to make platform and content available in containers. Before installing TrueSight Orchestration, Docker Engine must be installed on each host.|
|Docker Swarm||Containerized platform must be installed in a Docker swarm mode. In a swarm mode, Docker provides cluster management, orchestration, and support for service failover scenarios amongst many other features. In a swarm mode, you create a cluster of hosts, on which containers are deployed with component images.|
|Docker Registry||Docker uses a registry as a repository to store images. BMC recommends that you push the images provided in the installation bundle in a registry before starting the deployments. Storing images to a central registry ensures that the images are ready to be pulled on any host, making the installation process less cumbersome.|
|NFS Server||In a classic mode, a secondary CDP (High Availability CDP) can be installed that provides failover capabilities. In container mode, you need not install a secondary CDP as basic failover situations (such as services down or Docker Engine down) are handled by Docker Swarm. In addition to the swarm mode, for data persistence, you configure Network File System (NFS), which is a client-server based application that enables a shared file system over a network. You can use NFS with multi-node containerized platform environment, where the repository and CDP are on separate nodes.|
|NFS Client||You can designate more than one host or node as an NFS client. Client nodes must be in the swarm. Swarm ensures that the NFS client nodes are in a cluster.|
|External mounted volumes||In a Docker world, volumes are used for persisting data generated and used by the containers. Volumes are easier to manage and can be safely shared among containers. When you deploy the component images in your containers, TrueSight Orchestration mounts the volumes on the hosts. For example, for the repository, data from the tomcat logs, repository folder, and the config folders is retained in a an externally mounted volume. Volumes are managed by Docker and you do not need to perform any additional steps to configure volumes.|
|Remedy SSO in HA||In a high availability environment, multiple instances of the Remedy Single Sign-On web application are deployed and form a cluster. Usually a load balancer is used as a front end to the cluster, giving the external applications the appearance of a single server.|
|Repository HA||By default, the repository contains an embedded database and does not support high availability. You can enable high availability for the repository by configuring the repository to use an external database and leveraging the clustering features of the external database to provide failover.|
After understanding the high availability deployment, use the following table to get started with setting up an HA environment for TrueSight Orchestration in containers.
|Install Remedy Single Sign-On using HA clusters|
|Configure the repository to use an external database|
|Configure NFS for the CDP||Installing TrueSight Orchestration in containers in the high-availability mode|