Apaceh Hive Maintenance

Service Maintenance – plan and guide

Introduction

This document covers the maintenance activity plan and maintenance task that needs to carried out in the Hadoop Production cluster.

Maintenance activity schedule

The maintenance activity is planned for every fortnightly Sunday between 10 PM (Sunday) to 2 AM (Monday) EST.

*During this window there might be intermediate disruption in services.

Maintenance schedule - 2017

Month Date

Jan 15 29

Feb 12 26

Mar 12 26

Apr 9 23

May 7 21

Jun 4 18

Jul 2 16 30

Aug 13 27

Sep 10 24

Oct 8 22

Nov 5 19

Dec 3 17 31

Services maintenance

Hive service.

Service Restart

The hive service consists of two process one is the hive server and other is the hive metastore. These process needs to be restarted sequentially.

Pre Restart Check

Check if the hive server and metastore services are running. Make note of the PID and process time stamp.

Service restart sequence.

Using Ambari 2.3.2

1. Login into server where hive server and metastore are installed

2. Switch as hive user

sudo su - hive

3. Get the hive server PID and time

ps-ef | grep hiveserver

4. Go to Ambari->hive->summary->HiveServer2->Stop

5. Kill the hiveserver process PID and verify it’s not running

kill -9 <PID>

6. Get the Hive Metastore PID

ps -ef | grep hivemetastore

7. Go to Ambari->hive->summary->HiveMetastore->Stop

8. Kill the hivemetastore process PID and verify that it’s not running