Fall back in teradata is Data Protection mechanism at AMP level .It Provides high availability of data even in case of AMP Failure .If we enabled this option for a table , each and every row in a table has a fallback row in another AMP within the same cluster . If the main amp goes down still system perform operations by using duplicated or fallback rows in other amp.
Benefits of fallback:
- Access to rows incase AMP Failure.
- Adds extra level of data protection beyond disk array RAID1&5
- Highest level of data protection for teradata is RAID1 and fallback
- Automatically restores data changed during amp off-line
- It is critical for high availability applications.
Fallback cluster:
Cluster is nothing but group of AMPs. Cluster has equal number of AMPs for a group for entire system (Homogenous distribution of AMPs among the network). Here fallback is maintained in the same group of AMPS and in case of AMP failure it scans for remaining rows in Same Cluster (same group of AMPs).
In this way it avoids scanning of all AMPs and provides group scanning. If two AMPs in same group go down RDBMS halted. It is recommended to take 4 amps in a group since its average performance is good.
- A defined number of amps treated as fault tolerant unit
- Fallback rows for amps in a cluster resides in the same cluster.
- Loss of AMP in a cluster permits continued table access.
- Loss of two AMPs in a cluster leads to halt of RDBMS.