Today, we will discuss about “What is HA Implementation of Node” in Big data testing and what are the test-scenarios for HA testing and other points can be considered
In simple words, by creating one more standby replica node of the Working node and can be use if Working node downs due to any reason. If we have Replica node, this would ensure no down time and very minimum jobs failure in case hardware fails/any outage. Both nodes should be same as hardware, software and configuration wise except IP and Hostname are different to identify.
- Both Nodes should accessible and available in their respective status- one is active and running and other is on Passive mode.
- Both nodes have same connectivity, configurations from other systems i.e. S3 location, database connection etc.
- If switchover happens, Passive nodes switches (in few seconds) to Active and Active become Passive node till next switching.
- In any condition, both node’s status either Active or Passive.
- If Active node comes back, work as a Passive node.
- All configured jobs (of all users) should run on both nodes successfully.
- Discontinued job (Stopped due to Active node down) would not run from Active node after switchover happens
- Email notifications system should work from both nodes.
- If any monitoring/Nagios alert system available, should work on both nodes.
- Email functionality should implemented on both nodes for Switch-over operations.