Data Protection Overview
System Administration Manual
301
Buffering the requests in local NVRAM ensures that software failure or power failure will not
result in data loss. Additionally, nodes in a Titan cluster will mirror the contents of their NVRAM
to a partner node so that even a single server hardware failure will not result in data loss.
Power Failure Recovery
In the event of a power failure the contents of the NVRAM are preserved (using dedicated battery
backup) for a maximum of 72 hours. Once power is restored, the original server can roll back all
its File Systems and replay the modifications stored in the NVRAM log.
Buffering in a Cluster Configuration
When Titan is configured as a cluster, in addition to buffering all the File System modifications,
each Cluster Node mirrors the NVRAM contents of the other Cluster Node which ensures data
integrity in the event of a Cluster Node failure. When a Cluster Node takes over for the failed
node, it uses the contents of the NVRAM mirror to complete all File System modifications that
were not yet committed to the storage by the failed server.
NVRAM data is mirrored between the Cluster Nodes in this diagram.
FS NVRAM Statistics
Statistics
are available to monitor NVRAM activity. The statistics are updated every ten seconds.