Service availability and data integrity are key metrics for enterprises that run business critical applications at multiple remote locations. To ensure applications remain available, the infrastructure (servers and storage) which they run on, is grouped or “clustered” together, eliminating single points of failure within the infrastructure.
Clustering is primarily used to protect against unplanned downtime enabling redundant infrastructure components to take over the application/service workload, following a component failure. Equally, clustering can be used to minimize service disruption for planned downtime where IT operations are restricted, in order to perform upgrades, carry out repairs, etc.
To make clustering possible, some form of shared storage is required. SvSAN provides this capability using internal server storage, eliminating the need for expensive SAN switches and external storage arrays. SvSAN has its own clustering mechanism built-in, to ensure that the shared storage remains available while protecting the integrity of the data at all times.
This white paper describes the more common failure scenarios and explains the expected cluster behaviour during the failure and subsequent recovery.