High Availability (HA) Clustering is a means to ensure resilience within IT infrastructure. For business critical applications a HA architecture is key to ensuring that time and money are not lost to the failure of a single component system.
Modern infrastructure components offer HA functionality which can be incorporated into the design of a HA application stack including:
- Clustering of virtualisation hypervisors to protect against the failure of a single server with tools like RHEV
- Clustering of storage servers to protect against the failure of a single node with tools like Gluster, CEPH and Red Hat Enterprise Storage
- Deployments of Cloud Infrastructure-as-a-Service (IaaS) solutions such as OpenStack can be designed to eliminate single points of failure
In addition to these options many modern applications such as Oracle or MongoDB include their own clustering technology which can provide resiliance against component failure. For other use cases however it is still necessary to deploy standard Linux HA tools (Corosync, Pacemaker) to support the configuration of an application cluster and in these cases solutions from Open Source vendors such as Red Hat & SUSE can be an excellent fit.
With Linux HA multiple systems can be configured to operate as a single cluster. The features of Linux HA include:
- Resource agent handling for running services/middleware
- Resource failover
- HA Logical Volume Management
- Node fencing, a process where nodes not operating correctly are removed from the cluster
Securelinx is happy to offer solutions and implementation services for high availability on Linux, both in terms of infrastructure design and application led best practice, and the deployment and support of Linux HA Clusters - please contact us for more information.