Resolving Aster Software Issue on a Worker Node - Aster Analytics on AWS

Teradata Aster Analytics on AWS Getting Started Guide

Product
Aster Analytics on AWS
Release Number
1.5
Published
October 2017
Language
English (United States)
Last Update
2018-04-13
dita:mapPath
gkp1499384139306.ditamap
dita:ditavalPath
AWS.ditaval
dita:id
tsu1471542219751
lifecycle
previous
Product Category
Cloud
Follow the steps here when the Aster Software running on a worker node is having problems, but the node VM is fine.
  1. Remove the worker node from AMC. Log on to AMC, select Admin > Cluster Management and remove the worker node by clicking the X.
  2. Balance data across all available nodes. In AMC, select Admin > Cluster Management and click Balance Data.
  3. Terminate the worker node from AWS. Log on to AWS, select EC2, search for the instance and pick the worker node, then select Actions > Instance State > Terminate.
  4. Expand the number of nodes in AWS. Select EC2 > AUTO SCALING > Auto Scaling Groups, and select the affected cluster . On the Details page shown, update Desired, Min and Max to N+1, where N is the current number of nodes.
    Make sure that all the three parameters Desired, Min and Max are set to the same number.
    Wait until the newly added node is deployed and initialized in AWS.
  5. SSH to the queen node, load the Aster environment /home/beehive/config/asterenv.sh, and balance data across all available nodes.
    source /home/beehive/config/asterenv.sh
    > ncli system balancedata
  6. Verify the failover is complete.
    1. Make sure that all nodes are in active mode.
      > ncli node show
    2. Make sure the value in /home/beehive/config/initialPartitionCount is the same as that in /home/beehive/config/totalPartitionCount.
    3. In AWS, verify that the newly added node has the same network configuration (VPC, Availability Zone, Subnet, Security Group) as the existing nodes.
    4. Make sure you are be able to ssh on to the new node using the same key pair.
    5. Log on to AMC, in Nodes tab, verify that the new node has at least one primary partition.