To avoid triggering a node failure, don't stop or terminate instances before shutting down the database. When a node failure occurs on a single-node system, AWS starts up a replacement automatically unless there are issues with the operating system or in the database itself. To handle these situations, Teradata recommends deploying a new VantageCloud MPP system instance. Even if a node does not go down due to a failure, an instance still starts up after you stop or terminate an instance. Stopping instances does not delete attached EBS volumes, but terminating instances detaches and deletes the EBS volumes. Node failures are handled differently on local storage instance types.
To stop or terminate instances, you must first shut down the database.
- Stop the database.
# tpareset -x stop systemPutting the database in this state takes a few minutes. - Wait for the database to shut down.
# psh pdestate -w
PDE state: DOWN/HARDSTOP - From the AWS EC2 Console, terminate or stop the instances associated with the Teradata® VantageCloud Enterprise on AWS (DIY).
- If you are stopping an orphaned VantageCloud node in response to a node failure, do one of the following:
- From the AWS EC2 Console, terminate or stop the instances.
- From the command prompt, type psh init 0 to shut down the system.
Orphaned means when a failed node is recovered, but the recovery is too late for the node to rejoin the configuration and so the node is replaced. The node is online, but is no longer part of the system configuration.