Quantum Barcode Reader 3.5.1 User Manual

StorNext User’s Guide

349

Appendix B

B

HA Failover

StorNext is designed to be a resilient data management solution. StorNext 

supports operation in degraded mode and provides functionality to 

guarantee data protection in the event of a storage device failure or total 

site outage. For certain environments though, additional protection is 

required to deliver a higher level of availability. To meet these demands, 

StorNext includes MetaData Controller (MDC) failover. 
MDC failover allows a secondary MDC to take over StorNext operations 

in the event a primary MDC fails. Failover is supported for all StorNext 

management operations including client IO requests (File System) and 

data mover operations (Storage Manager). MDCs in a failover pair 

typically run in an active / passive configuration, but both MDCs can be 

configured to run active File System processes. In the event one MDC 

fails, the other continues to perform its current operations, as well as 

those of the failed MDC.

Like all failover solutions, StorNext must provide functionality to prevent 

a damaged or inaccessible MDC from incorrectly processing IO requests 

that should be handled by the active MDC (often referred to as a “split 

brain” scenario). To handle this, StorNext utilizes a special failover 

methodology call STONITH - shoot the other node in the head. STONITH 

Active / Active Storage Manager processes are not currently 

supported in MDC failover.