High Availability

ClustrixDB nResiliency

The patented ClustrixDB nResiliency technology allows you to prepare your Clustrix cluster to be fault tolerant up to n number of nodes by simply setting a parameter. The Clustrix Rebalancer will optimize the database setup and data distribution to accommodate a catastrophic failure of a set number of nodes, and in the event a failure of that magnitude occurs, will gracefully handle and replace the failed nodes while maintaining availability.

Automatic High Availability

By nature, the cloud is a moving substrate to rest your data, so we built ClustrixDB to automatically accommodate for this, and protect your data availability—always ensuring there are at least two copies of your data, even with a node outage. No intervention is needed, the system is designed to work this way for you. This is quite the difference from other systems, which would typically need some level of administrator intervention to promote a read slave, and likely incur some level of lag and inconsistency from the previous master’s transactional state.

Transparent To Developers

ClustrixDB nResiliency is the first in the market to offer complete confidence that your valuable data is safe and continuously available should two or more servers (nodes) fail at the same time. ClustrixDB then automatically generates the number of data replicas (and Paxos Acceptors) necessary to successfully recover, in the event of multi-node failure. And since data locality is all transparent and optimized for developers already, no application changes are ever necessary—even in the case of multi-node failures.

Configurable Multi-Node Fault Tolerance

ClustrixDB’s nResiliency feature provides the ability to define the the number of nodes in the cluster that can become unavailable simultaneously while ensuring continuous database availability. The number of server redundancy is a factor of cluster size and Paxos ‘split-brain’ math. The maximum number of redundancy is ‘number of cluster nodes’ divided by two, and minus 1. For example:

 

nResiliency can be used to architect your system for a variety of fault tolerance requirements:

  • Set MAX_FAILURES at a high number for their high-value data that are necessary to keep mission-critical applications running in the event of simultaneous failures
  • Set MAX_FAILURES at a mid-range number for high volume data that are not required to have multiple levels of redundancy
  • Set MAX_FAILURES at a low number for high-throughput, ‘fast-lane’ data which can be easily replaced

Disaster Recovery (DR)

ClustrixDB fully supports MySQL replication, parallelizing it via our patented technology. ClustrixDB supports Master/Slave, Master/Master, active/passive, and active/active replication between clusters. ClustrixDB can support MySQL slaves, as well as be a slave to a MySQL master.

ClustrixDB clusters can be deployed across geographical regions with asynchronous replication. For disaster recovery Clustrix offers fast parallel backup, streaming directly from each server, resulting in up to 10x faster than an equivalent MySQL backup.