ClusterLabs
The open-source high-availability cluster stack
The ClusterLabs community provides a collection of free and open-source software projects related to high-availability clustering, including Corosync and Pacemaker.
Benefits
A high-availability cluster coordinates the start-up, monitoring, and recovery of services across a set of host machines, to minimize downtime while preserving data integrity.
Unlike a local boot system such as systemd, a ClusterLabs cluster is aware of the context of all hosts in the cluster, allowing recovery from node-level failures as well as intricate relationships between services across different hosts.
ClusterLabs clusters support fencing, such as powering off unresponsive nodes, to ensure data integrity and availability after host failure. Without fencing, hosts that lose communication can corrupt data by incorrectly bringing up services on more than one host.
Features
- Extensive configuration options to control cluster behavior
- Detection and recovery from host- and service-level failures
- Monitoring any measure of node health
- Practically any host redundancy configuration
- Multi-site clusters
- Live migration of services that are capable of it, including virtual machines
- Supports services that should be active on more than one host at the same time
- Supports services that can be promoted into a special role, such as a database primary
- Supports containerized services including networking and storage requirements of multiple replicas
- Supports complex service relationships, including ordering and colocation, that can span multiple hosts