How to Remove Non-Critical Nodes from a BDA Cluster
(Doc ID 2244458.1)
Last updated on MAY 27, 2019
Applies to:Big Data Appliance Integrated Software - Version 4.6.0 and later
How to remove a non-critical nodes from a BDA Cluster. This applies to BDA V4.4 and higher.
This note also includes a section on how to add a non-critical node back into the cluster.
ACS, system admin, or anyone removing a non-critical node from a BDA cluster.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document
|Frequently Asked Questions|
|Where can questions on the Mammoth decommission step with bdacli be found?|
|Does removing or adding slave nodes to a cluster require a cluster restart or cluster downtime?|
|After removing a non-critical node out of the BDA cluster for use as an edge node, is it supported to change the disk mounts from /u01 to /u12 to be configured in RAID1?|
|When removing a non-critical node from a BDA cluster does /opt/oracle/bda/rack-network.json or /opt/oracle/bda/cluster-network.json require update?|
|When removing a non-critical node from a cluster why isn't it necessary to update /opt/oracle/bda/cluster-network.json to reflect the change?|
|Overview of the Steps to Remove a Non-Critical Node|
|Detailed Steps to Remove a Non-Critical Node|
|Make sure the cluster is healthy|
|Look for Under-Replicated Blocks and Set the Replication Factor to 3|
|Decommission the Node and Ensure the Cluster is Healthy|
|Stop/Delete All Roles on the Host in Cloudera Manager|
|Delete the Host from Cloudera Manager|
|Restart Services/Deploy Client Configuration so the Cluster is Healthy|
|Update /opt/oracle/bda/install/state/config.json and /opt/oracle/bda/cluster-hosts-infiniband to remove all references to the removed node.|
|Ensure the Cluster is Healthy|
|Steps Required for Clusters Running Big Data SQL 3.1.|
|Run a HDFS Rebalance|
|How to Add a Non-Critical Node Back into the Cluster|