Shorten Planned and Unplanned Downtime
Keeping your data intact and your important applications up and running 24×7 matters. InterSystems IRIS provides several options for high availability (HA) and disaster recovery (DR), including clustering, virtualization HA, and an elegant, easy-to-implement technology for database mirroring.
A database mirror is a logical grouping of two InterSystems IRIS systems. Upon startup, the mirror automatically designates one of these two physically independent systems as the primary system; the other one automatically becomes the backup system. Mirrored databases are synchronized from the primary to the backup failover member in real time through a TCP channel.
Sharded database architectures require setting up a database mirror for each shard, thereby eliminating any single point of failure. Deploying in a cloud environment will require some extra configuration steps to ensure automatic redirection of incoming traffic to the primary node.
With database mirroring, application recovery time is typically reduced to seconds. The use of mirroring also enables minimal (or even zero) downtime upgrades.
Using Database Mirroring for Disaster Recovery
An asynchronous mirror member can be set up at a remote site, and updated in near real time. If the primary data center fails, your data will not be lost. Disaster recovery when both members are deployed in a public cloud is dependent on the provider capabilities, but can be achieved by setting up asynchronous members in different “regions,” or even between clouds from different providers.
Clustering and Virtualization
Clustered systems are typically dependent on shared disk access, but with only one system active at a time. If the active system fails, InterSystems IRIS is automatically started on another server that takes over the processing responsibilities. Users must sign back on to the new server, which may cause a noticeable delay. Virtualization HA works in much the same way.