Sample script to Onboard and run Hive queries from multiple users

mutipleuser_query.sh

#onboard N users and HDFS home directories

numusers=5
userprefix="testuser"
group="testusers"
users=()
tables=("hortoniabank.ww_customers" "hortoniabank.us_customers" "finance.tax_2009" "finance.tax_2010" "finance.tax_2015" "cost_savings.claim_savings" "claim.provider_summary" "consent_master.consent_data")
export hive_port=10500

groupadd ${group}

for i in $(seq 1 ${numusers})
    do
      user=${userprefix}${i}
      echo "Creating user ${user}"
      useradd -g ${group} ${user}   #must be done on all compute nodes!

      echo "Creating home dir for ${user}"      
      sudo -u hdfs hdfs dfs -mkdir /user/${user}
      
      users+=(${user})
    done
    
    
echo "${users[*]}"


#Simulate random user querying a random table at random time

# Seed random generator
RANDOM=$$$(date +%s)


while [ 1 ]
do
    # Get random user...
    selecteduser=${users[$RANDOM % ${#users[@]} ]}

    # Get random table...
    selectedtable=${tables[$RANDOM % ${#tables[@]} ]}

    beeline_url="jdbc:hive2://localhost:${hive_port}/default"
    
    # Run query as random user
    echo "$selecteduser: select count(*) from $selectedtable"
    beeline -n $selecteduser -u ${beeline_url} -e "select count(*) from $selectedtable"

	#sleep for random time between 1-10s
    randomsleep=$(( ( RANDOM % 10 )  + 1 ))
    sleep ${randomsleep}
done