AggreGate has built-in support for building a failover cluster to ensure high availability of provided services. The failover cluster consists of two or more installations of AggreGate Server and one or more installations of the underlying database engine.
|
High availability is a system design approach and associated service implementation that ensures a prearranged level of operational performance will be met during a contractual measurement period. |
The objective of failover cluster is reaching 100% availability of the server. The cluster consists of Master Node and one or more Failover Nodes. During normal operation, the master node is servicing all operations. All failover nodes are running in standby mode and monitoring the state of master node.
The failover nodes are automatically switched to Failover Master mode (i.e. activated) in the following cases:
| • | Network or power outage of the master node |
| • | Major hardware or software failure of the master node |
Both master and failover nodes are full installations of AggreGate Server that can work on its own. However, master and all failover nodes share the same database.
AggreGate Cluster Features
The integrated failover clustering engine has some unique features:
| • | No dependencies to third-party software or operating system services, such as Linux Heartbeat or Microsoft Cluster Service |
| • | Cluster nodes intercommunication is performed via the database, no additional setup of node-to-node IP communications is necessary |
| • | Database mirroring is optional and may be implemented using DBMS native replication or AggreGate-based replication |
| • | Cluster nodes benefit from database load balancing |
| • | Cluster nodes may run on different operating systems and hardware |
Failover Cluster Basics
The failover cluster includes two separate levels of ensuring high availability service: database failover and AggreGate Server failover.
Database failover implies setting up more than one installation of the database engine. This database engine is used by both master and failover nodes. If one of the database installations fail, the other(s) will continue to server AggreGate Server requests.
AggreGate Server failover technology comprises one Master server and one or more Failover servers that share the same (possibly clustered/replicated) database. The failover servers are activated when the master server fails due to any reason.
|
AggreGate Master server may run on the same physical machine with the "first part" of clustered/replicated database installation, while Failover server may share physical hardware with the second database engine installation. This allows to build a complete failover solution using just two physical servers. |
For further information see:
Failover Cluster Setup
Perform the following steps to set up AggreGate Server failover cluster:
| • | Install two or more copies of AggreGate Server on different physical servers. |
| • | Configure database failover on every server, i.e. configure all servers to use a single shared (and possibly clustered/replicated) database. In most cases it will be sufficient to complete AggreGate-based database replication configuration procedure. |
| • | Change Cluster Role global configuration setting of the one server to Master. Change roles of the other servers to Failover. See Server Configuration for details. |
| • | Change Failover Mode of one failover server to the Normal. Change mode of the other failover servers to Read-only. |
| • | Start all servers. |
|
Failover servers must be started in Service Mode. |
The master node should start normal operations, while failover nodes should switch to standby mode, showing "Monitoring Master Server Status" message on the splash screens.
Cluster Servers Security
All servers in the cluster have access to all information flowing inside AggreGate. Therefore, the same security precautions should be taken for both Master and Failover servers.