Remove a node
Remove an unresponsive node
If a VM or machine in a universe reaches its end of life and has unrecoverable hardware or other system (OS, disk, etc.) issues, it will be detected and displayed in the Yugabyte Platform console as an
Unreachable node. The Raft consensus algorithm will ensure that other leaders will get elected for the underlying data shards. But, the universe is in a partially under-replicated scenario and will not be able to tolerate many more failures. So a quick remedy is needed.
So, for the
yb-15-node-actions-n3 node, you can use the Actions drop-down list at the end and choose the Remove Node option. This will bring the node to the
Removed status and no further data can be accessed from this node as server processes are shut down.
Note that there is no
MASTER/TSERVER shown for that node. If that node was the
Master Leader, then a Raft election will move it to another node quickly. Similar leader elections happen for tablets for which this node was the the leader tablet server.
Release a node instance
Because you know that the instance is dead, you can go ahead and release the IP address using the Actions drop-down list to select the Release Instance option at the end of the Removed node. It will show up as a