Symfoware Server supports 1:1 standby.
Symfoware Server supports failover operation of 1:1 standby by using the Symfoware Server Hot Standby Option or Symfoware Server Parallel Cluster Option.
Symfoware Server Parallel Cluster Option enables 1:1 standby load-shared operation.
For details, see the "SymfoWARE (R) Server Cluster Installation/Administration Guide for PRIMECLUSTER."
Because Symfoware Server realizes failover, you can build a high-availability system. The operation involved in performing failover is referred to as the failover operation.
Failover refers to the facility that enables the taking over of a job that is being operated on the relevant node to another node (standby system) if an error occurs on one of the nodes in the cluster system. The following two functions of Symfoware Server are used to realize failover:
Standby function
Hot standby function
The standby function is such that, if an error occurs in a node, the standby RDB system to which the operation is to be switched is started up to take over the job operation.
To cope with a node error, the RDB system on the standby system is started up in advance and, if an error occurs on the node, failover of the operation can be completed in a short time. This is called the hot standby function.
Symfoware Server realizes load sharing to support high scalability. Load sharing refers to the function to process transactions in parallel on multiple nodes in the cluster system. The operation for performing load sharing is referred to as the load-shared operation.
Load sharing has the following two features:
Performance on a par with a high-reliability system can be realized.
One of the features of the load-shared operation is that you can realize a level of performance that is comparable with that of a high-reliability system.
A load-shared operation is controlled by PRIMECLUSTER. The purpose of this is to ensure compatibility with other middleware products included in the cluster system by entrusting cluster system management and operations to cluster control. You can thus make maximum use of the high-scalability and high-reliability that are realized when cluster control is combined with the hardware
Load-shared operation enables communication between the nodes that constitute the cluster system.
An entire system can be realized as a high-reliability system.
With load-shared operation, you can restart a job within a few seconds of a node failing.
With load-shared operation, database updated logs are transmitted sequentially, during operation, into memory of the standby node by using TCP/IP or BLASTBAND. If a switchover event occurs, the flash treatment recovery will be instigated and the first data will be closed by using the updated logs on the standby node. This enables the restart of the job transaction.
Failure recovery for lock-out data will be performed in the background in parallel with the job transaction.
