failover cluster
Failover stands for fail-safe. A failover cluster is a cluster that provides fail-safe services. Such failover clusters have a redundant structure and guarantee the fail-safety of web servers, storage systems or databases. In case of a server or storage system failure, the parallel running cluster takes over the continuous data storage by replication.
A fault-tolerant failover cluster consists of two synchronously operating backup servers, one of which is active while the second operates in standby and monitors the active one. If the active server fails, the standby server takes over its function. The handover is performed by the cluster manager or another central management component. It takes a certain amount of time, which is the failover time, during which the application data is not available.
From the logic of data replication, failover systems can hook into the input/output system of the operating system or the software component to be backed up and take over the backup operations. There are also different concepts for the actual switchover. With automatic failover, the replicating server (standby) regularly queries its active master and takes over data replication as soon as it cannot reach the master. It can also trigger a switchover if a switch fails or a network is temporarily overloaded. For the regular polling of the master, the replication server uses a heartbeat, which is a control signal with which the replication serverchecks the functionality of the master.
Failover clusters are characterized by high fault tolerance and can tolerate errors even if several errors occur simultaneously. These systems operate fault-tolerantly until no further redundant components are available to protect the storage system against further faults.