Failover Process - Teradata Data Mover

Teradata Data Mover User Guide

Product
Teradata Data Mover
Release Number
16.10
Published
June 2017
Language
English (United States)
Last Update
2018-03-29
dita:mapPath
kmo1482331935137.ditamap
dita:ditavalPath
ft:empty
dita:id
B035-4101
lifecycle
previous
Product Category
Analytical Ecosystem

The dmcluster service script is used to run various commands on the cluster. The dmcluster service script can be found in /opt/teradata/client/nn.nn/datamover/failover/, where nn.nn is the major and minor version numbers of Data Mover, for example 15.11. After the cluster configuration completes using dmcluster config, all master components are started in master mode. If the master daemon, agents or sync service are running, they are stopped and restarted in master mode. All slave components are stopped and the sync service on the remote sync server is started in slave mode. The monitoring service starts monitoring the master components by making SSH connection to the master component servers at regular intervals to check if the services are up and running. If the master Data Mover daemon, master repository, or all agents used by the master daemon are down, a failover sequence is initiated on the slave daemon server. The failover sequence shuts down the daemon, agents and sync service on the master servers and starts the daemon, agents and sync service on the standby servers in master mode. An attempt is made to the start the old master sync service in slave mode.

Once the failover sequence completes, the monitoring service shifts to the standby monitoring server and begins monitoring the new master components. If the monitoring service finds a running Data Mover daemon process on the slave daemon server, it will automatically shut it down to prevent two daemons from being active at the same time.

Running the dmcluster status command should at this point should show the following:
  • The Data Mover daemon is NOT RUNNING on local.daemon.host
  • The Data Mover sync service is RUNNING as slave on local.repository.host
  • The Data Mover agents are be NOT RUNNING on the hosts specified for local.agents.host
  • The monitoring service is NOT RUNNING on local.monitor.host
  • The Data Mover daemon is RUNNING on remote.daemon.host
  • The Data Mover sync service is RUNNING as master on remote.repository.host
  • The Data Mover agents are RUNNING on the hosts specified for remote.agents.host (unless any agents are shared by master and slave daemon)
  • The monitoring service is RUNNING on remote.monitor.host

The dmFailover.log provides more information on the failover sequence.