The impact of offline AMPs on the Update driver depends on:
- The number of offline AMPs in a cluster, either logically or physically
- The operational phase of the Update tasks when the offline AMP condition occurs
- Whether the target tables are fallback or nonfallback
The table below describes the impact of offline AMPs on Update driver tasks on fallback and nonfallback tables.
|All of the target tables are fallback
|Not more than one AMP is offline
|Update driver tasks continue to run
as long as there is not more than one AMP offline in a cluster,
either logically or physically.
The offline AMP does not participate in the application phase if:
|Two or more AMPs are offline
|Update driver tasks do not run or
terminate if two or more AMPs are offline in a cluster, either
logically or physically.
In the application phase, if AMPs are offline to the extent that data on the disk is corrupted, then you must restore the affected tables.
|One or more of the target tables is nonfallback
|One or more AMPs are offline
|Update driver tasks terminate and
you cannot restart until all of the AMPs are online.
The Update driver also terminates if I/O errors corrupt the target tables in the application phase.