Skip to content

Instantly share code, notes, and snippets.

@ashrithr
Created March 8, 2015 22:05
Show Gist options
  • Star 4 You must be signed in to star a gist
  • Fork 8 You must be signed in to fork a gist
  • Save ashrithr/1d2c71f0ef2663ebe6e7 to your computer and use it in GitHub Desktop.
Save ashrithr/1d2c71f0ef2663ebe6e7 to your computer and use it in GitHub Desktop.
YCSB stress test hbase commands
#!/usr/bin/env bash
#
# Simulates mixed workload on HBase using YCSB
# Author: Ashrith (ashrith at cloudwick dot com)
# Date: Wed, 16 2014
#
#
# You may want to tweak these variables to change the workload's behavior
#
# Number of total rows to insert into hbase
RECORDS_TO_INSERT=100000
# Total operations to perform for a specified workload
TOTAL_OPERATIONS_PER_WORKLOAD=10000
# Throttling (specifies number of operations to perform per sec by ycsb)
OPERATIONS_PER_SEC=1000
# Number of threads to use in each workload
THREADS=5
# Name of the hbase column family
HBASE_CFM="f1"
# Number of hbase regions to create initially while creating the table in hbase
HBASE_RC=16
# Log file to use
LOG="/tmp/hbase_ycsb.log"
#
# NOTE: DON'T CHANGE BEYOND THIS POINT SCRIPT MAY BREAK
#
# Create a table with specfied regions and with one column family
echo "Creating pre-splitted table"
hbase org.apache.hadoop.hbase.util.RegionSplitter usertable HexStringSplit \
-c ${HBASE_RC} -f ${HBASE_CFM} >> $LOG
# Load the dataset (no throttling)
echo "Loading intial dataset"
bin/ycsb load hbase -P workloads/workloada -p columnfamily=${HBASE_CFM} \
-p recordcount=${RECORDS_TO_INSERT} \
-s -threads ${THREADS} >> /tmp/hbase_load.log
#
# => Workload simulation (Throtlling 10000 operations/second)
#
echo "Simulating workloads"
# Update heavy workload Read/Update ratio: 50/50.
bin/ycsb run hbase -P workloads/workloada -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloada.log
# Read mostly workload Read/Update ratio: 95/5
bin/ycsb run hbase -P workloads/workloadb -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloadb.log
# Read Only workload Read/Update ratio: 100/0
bin/ycsb run hbase -P workloads/workloadc -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloadb.log
# Read Modify Write workload Read/Update ratio: 50/50
bin/ycsb run hbase -P workloads/workloadf -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloadb.log
# Read latest workload Read/Update/Insert ration: 95/0/5
bin/ycsb run hbase -P workloads/workloadd -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloadb.log
# Short ranges workload Scan/insert ratio: 95/5
bin/ycsb run hbase -P workloads/workloade -p columnfamily=${HBASE_CFM} \
-p operationcount=${TOTAL_OPERATIONS_PER_WORKLOAD} \
-target ${OPERATIONS_PER_SEC} \
-threads ${THREADS} -s >> /tmp/workloadb.log
# Count rows in a table
# echo "count 'usertable'" | hbase shell
# Delete the table contents
# echo "truncate 'usertable'" | hbase shell
# Disable and Drop table existing
echo "Disabling and dropping table"
echo "disable 'usertable'" | hbase shell >> $LOG
echo "drop 'usertable'" | hbase shell >> $LOG
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment