Requirement:
Aruba IntroSpect Analyzer running 2.5 or older version
Applies for both Hardware and AWS Analyzer instances
Solution:The shutdown process is very simple and we recommend logging on the shell on Analyzer and run the command to shutdown the Analyzer.
Once the node is powered back, we need to start the services manually either from Analyzer Cli or UI.
Configuration:Shut Down Process:
In order to shutdown the Analyzer, please run the below command on Analyzer cli:
analyzer-manager --stop --shutdown-linux --continue-on-error
When it’s completed, all the nodes are powered off and ready to pull the power if required in case if you are moving the Appliances.
Power Up Process:
Once the node is powered back, we need to start the services manually from Analyzer Cli:
analyzer-manager --start -c
VerificationShut Down Process:
[root@ ~]# analyzer-manager --stop --shutdown-linux --continue-on-error
This will stop all the Niara Analyzer services on all the hosts,
including all Hadoop services.
Continue? [y/N] y
Shutdown option will run linux shutdown command on all the Niara Analyzer cluster nodes
after all the services are stopped for safe power off.
Continue? [y/N] y
Stopping the analyzer service on the an-node host...
... Ok
Stopping the analyzer_sched service on the an-node host...
... Ok
Waiting for pending YARN jobs to run on the cluster...
No YARN Application running; confirming...
No YARN Application running; confirming...
... Ok
... Ok
Stopping the spark-history-server service on the yarnrm host...
... Ok
Stopping the hbase-thrift-server service on the hbasethrift hosts...
... Ok
Stopping the opentsdb service on the opentsdb hosts...
... Ok
Setting cluster maintenance mode to enabled...
... Ok
... Ok
Stopping all of the Hadoop services via the manager
... Ok
Stopping the management service...
... Ok
Stopping the redis service on the an-node host...
... Ok
Stopping the elasticsearch service on the es hosts...
... Ok
Turning on the LEDs on the cluster hosts...
... Ok
About to run poweroff command on all of the Hadoop nodes...
... Ok
About to run poweroff command on Analyzer node...
Broadcast message from root@is-tac.arubans.us
(/dev/pts/2) at 16:26 ...
The system is going down for power off NOW!
Power Up Process:
[root@ ~]# analyzer-manager --start -c
Starting the elasticsearch service on the es hosts...
... Ok
Starting the redis service on the an-node host...
... Ok
Starting the management service...
... Ok
Waiting 60 seconds for the management service...
Checking that all hosts are registered with the manager...
... Ok
Checking host status with the manager...
... Ok
Running the host inspector...
... Ok
Starting all of the Hadoop services via the manager:
Starting ZOOKEEPER...
... Ok
Starting HDFS...
... Ok
Starting HBASE...
... Ok
Starting TEZ...
... Ok
Starting KERBEROS...
... Ok
Starting MAPREDUCE2...
... Ok
Starting YARN...
... Ok
Starting SPARK...
... Ok
Starting FLUME...
... Ok
Starting the opentsdb service on the opentsdb hosts...
... Ok
Starting the spark-history-server service on the yarnrm host...
... Ok
Starting the hbase-thrift-server service on the hbasethrift hosts...
... Ok
Starting the analyzer_sched service on the an-node host...
... Ok
Starting the analyzer service on the an-node host...
... Ok
Setting cluster maintenance mode to disabled...
... Ok
Manager summary: All services started as expected
Turning off the LEDs on the cluster hosts...
... Ok
analyzer-manager action completed.
Check the system health:
Execute the below command on Analyzer cli to check the status of all cluster services running on Analyzer:
[root@ ~]# analyzer-manager -d
Checking if the ambari-agent service is running on the cluster hosts...
... Ok
Checking if the ambari-server service is running on the an-node host...
... Ok
Checking if the ambari-server-db service is running on the an-node host...
... Ok
Checking the management services...
AMBARI_METRICS STARTED
...GOOD
... Ok
Checking the cluster services...
FLUME STARTED
...GOOD
HBASE STARTED
...GOOD
HDFS STARTED
...GOOD
MAPREDUCE2 STARTED
...GOOD
SPARK STARTED
...GOOD
YARN STARTED
...GOOD
ZOOKEEPER STARTED
...GOOD
... Ok
Checking if the opentsdb service is running on the opentsdb hosts...
... Ok
Checking if the collectd service is running on the cluster hosts...
... Ok
Checking if the analyzer service is running on the an-node host...
... Ok
Checking if the analyzer_sched service is running on the an-node host...
... Ok
Checking if the elasticsearch service is running on the es hosts...
... Ok
Checking if the redis service is running on the an-node host...
... Ok
Checking the Elasticsearch cluster status via host cdh-3...
epoch timestamp cluster status node.total node.data shards pri relo init unassign pending_tasks max_task_wait_time active_shards_percent
1556324191 00:16:31 cluster-internal-elasticsearch green 3 3 18 9 0 0 0 0 - 100.0%
... Ok
Checking if the spark-history-server service is running on the yarnrm host...
... Ok
Checking if the hbase-thrift-server service is running on the hbasethrift hosts...
... Ok
[root@ ~]#