The following terms are used throughout this document and, while some may be familiar to you, it may be helpful to review how SIOS defines and uses these terms.
Network Communication Terms
Crossover cable – A cable used to directly connect computing devices together, instead of being connected to a network switch, hub or router. This cable creates an isolated, private network to allow cluster-related and data replication traffic to flow between systems.
Types of LifeKeeper Servers
Server – A computer system dedicated to running software application programs.
Active Server – This is the server where the resource hierarchy is currently running (IN SERVICE).
Standby Server – This is the server where the resource hierarchy is defined, but is not currently running. This server is available to bring the resource hierarchy into service should something happen to the resource hierarchy on the Active Server.
Primary Server – This is the server in a LifeKeeper configuration with the highest priority for a given resource hierarchy. It is the server that provides services for the resource hierarchy under normal circumstances.
Secondary Server – This is the server in a LifeKeeper configuration with the 2nd highest priority for a given resource hierarchy.
Source Server – In a LifeKeeper cluster, using data replication, this is the Active Server. It is where the resource hierarchy is currently running and the replicated partition (Source Partition) is accessible for writes.
Target Server – In a LifeKeeper cluster, using data replication, this is the Standby Server. The replicated partition (Target Partition) is updated with writes from the Source Partition by the SIOS Data Replication system. This partition should not be accessed/modified manually.
SIOS Data Replication Terms
Replication – Transferring data from one partition to another via a sector-by-sector copy. During replication, the target partition should not be accessed or modified assuring your data integrity.
Synchronous – A replication scheme in which the data is confirmed written and valid on the target before the write operation occurs on the source disk through a series of information exchanges. Synchronous mirrors should only be implemented on high speed (100Mbps+) networks due to the network overhead involved.
Asynchronous – A replication scheme in which the data is released for writing on the source immediately and is sent to the target(s) simultaneously for writing as fast as the data can get there and can be written on them.
Rate of Change – A measure of the amount of data which is changing over a set period of time.
Compression – An algorithm which is optionally implemented to reduce the amount of traffic between source and target nodes. Nine levels of compression are offered. Compression is turned off by default.
Throttling – An optionally implemented mechanism to limit the bandwidth used for replication.
LifeKeeper Product Terms
Communications Path – A mechanism supporting communication between nodes in a LifeKeeper cluster. SIOS highly recommends implementing multiple communication paths between all servers in the cluster to eliminate a single point of failure.
Heartbeat – A periodic message exchanged between nodes in a LifeKeeper cluster that provides server health monitoring. A heartbeat message is one type of inter-node cluster communication sent over a communications path.
Split Brain – A situation in which all communications paths between cluster members fail, but all servers remain up and running. In this situation, both systems believe the other has failed and both believe they should keep or bring resources into service.
Failover – The unplanned migration of a resource hierarchy to the Standby Server because of a system or resource failure on the Active Server.
Switchover – The planned migration of a resource hierarchy from the Active Server to the Standby Server.
Switchback – The setting that governs the recovery behavior of the server where the resource was in service when it failed. If the setting is “intelligent”, the server acts as a possible backup for the given resource. If the setting is “Automatic”, the server actively attempts to re-acquire the resource without further notice.
Resource – A system asset that can be protected by LifeKeeper. Resources can be used to represent disk partitions, virtual IP addresses, applications, etc.
Extend a Resource – Create or define an already configured LifeKeeper resource onto another server in the cluster and build an equivalency relationship that prevents the resource from coming in service on both systems simultaneously.
Resource Hierarchy – A grouping of resources, in a predetermined order, from high to low. This may also be referred to as simply a Hierarchy.
Shared Storage – One or more logical disk partitions that are physically attached to all nodes in a cluster. LifeKeeper ensures that the volume is only accessible by one server at a time. This is formally called I/O fencing.
Data Replication (Disk Mirroring) – The replication of logical disk partitions to separate physical hard disks in real time to ensure continuous availability, currency and accuracy of data.
Source – The partition on the source server used for replication. The “gold” copy of the data.
Target – The partition on the target server used for replication.
Switchable IP Address – A unique IP address that may be moved between systems in the cluster. Client systems connect to this address and the system where the virtual IP resource is active will respond to requests.
このトピックへフィードバック