mxtasking/workloads_specification/workload_template

# Copyright (c) 2012-2016 YCSB contributors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you
# may not use this file except in compliance with the License. You
# may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied. See the License for the specific language governing
# permissions and limitations under the License. See accompanying
# LICENSE file.

# Yahoo! Cloud System Benchmark
# Workload Template: Default Values
#
# File contains all properties that can be set to define a
# YCSB session. All properties are set to their default
# value if one exists. If not, the property is commented
# out. When a property has a finite number of settings,
# the default is enabled and the alternates are shown in
# comments below it.
# 
# Use of most explained through comments in Client.java or 
# CoreWorkload.java or on the YCSB wiki page:
# https://github.com/brianfrankcooper/YCSB/wiki/Core-Properties

# The name of the workload class to use
workload=com.yahoo.ycsb.workloads.CoreWorkload

# There is no default setting for recordcount but it is
# required to be set.
# The number of records in the table to be inserted in
# the load phase or the number of records already in the 
# table before the run phase.
recordcount=1000000

# There is no default setting for operationcount but it is
# required to be set.
# The number of operations to use during the run phase.
operationcount=3000000

# The number of insertions to do, if different from recordcount.
# Used with insertstart to grow an existing table.
#insertcount=

# The offset of the first insertion
insertstart=0

# The number of fields in a record
fieldcount=10

# The size of each field (in bytes)
fieldlength=100

# Should read all fields
readallfields=true

# Should write all fields on update
writeallfields=false

# The distribution used to choose the length of a field
fieldlengthdistribution=constant
#fieldlengthdistribution=uniform
#fieldlengthdistribution=zipfian

# What proportion of operations are reads
readproportion=0.95

# What proportion of operations are updates
updateproportion=0.05

# What proportion of operations are inserts
insertproportion=0

# What proportion of operations read then modify a record
readmodifywriteproportion=0

# What proportion of operations are scans
scanproportion=0

# On a single scan, the maximum number of records to access
maxscanlength=1000

# The distribution used to choose the number of records to access on a scan
scanlengthdistribution=uniform
#scanlengthdistribution=zipfian

# Should records be inserted in order or pseudo-randomly
insertorder=hashed
#insertorder=ordered

# The distribution of requests across the keyspace
requestdistribution=zipfian
#requestdistribution=uniform
#requestdistribution=latest

# Percentage of data items that constitute the hot set
hotspotdatafraction=0.2

# Percentage of operations that access the hot set
hotspotopnfraction=0.8

# Maximum execution time in seconds
#maxexecutiontime= 

# The name of the database table to run queries against
table=usertable

# The column family of fields (required by some databases)
#columnfamily=

# How the latency measurements are presented
measurementtype=histogram
#measurementtype=timeseries
#measurementtype=raw
# When measurementtype is set to raw, measurements will be output
# as RAW datapoints in the following csv format:
# "operation, timestamp of the measurement, latency in us"
#
# Raw datapoints are collected in-memory while the test is running. Each
# data point consumes about 50 bytes (including java object overhead).
# For a typical run of 1 million to 10 million operations, this should
# fit into memory most of the time. If you plan to do 100s of millions of
# operations per run, consider provisioning a machine with larger RAM when using
# the RAW measurement type, or split the run into multiple runs.
#
# Optionally, you can specify an output file to save raw datapoints.
# Otherwise, raw datapoints will be written to stdout.
# The output file will be appended to if it already exists, otherwise
# a new output file will be created.
#measurement.raw.output_file = /tmp/your_output_file_for_this_run

# JVM Reporting.
#
# Measure JVM information over time including GC counts, max and min memory
# used, max and min thread counts, max and min system load and others. This
# setting must be enabled in conjunction with the "-s" flag to run the status
# thread. Every "status.interval", the status thread will capture JVM 
# statistics and record the results. At the end of the run, max and mins will
# be recorded.
# measurement.trackjvm = false

# The range of latencies to track in the histogram (milliseconds)
histogram.buckets=1000

# Granularity for time series (in milliseconds)
timeseries.granularity=1000

# Latency reporting.
#
# YCSB records latency of failed operations separately from successful ones.
# Latency of all OK operations will be reported under their operation name,
# such as [READ], [UPDATE], etc.
#
# For failed operations:
# By default we don't track latency numbers of specific error status.
# We just report latency of all failed operation under one measurement name
# such as [READ-FAILED]. But optionally, user can configure to have either:
# 1. Record and report latency for each and every error status code by
#    setting reportLatencyForEachError to true, or
# 2. Record and report latency for a select set of error status codes by
#    providing a CSV list of Status codes via the "latencytrackederrors"
#    property.
# reportlatencyforeacherror=false
# latencytrackederrors="<comma separated strings of error codes>"

# Insertion error retry for the core workload.
#
# By default, the YCSB core workload does not retry any operations.
# However, during the load process, if any insertion fails, the entire
# load process is terminated.
# If a user desires to have more robust behavior during this phase, they can
# enable retry for insertion by setting the following property to a positive
# number.
# core_workload_insertion_retry_limit = 0
#
# the following number controls the interval between retries (in seconds):
# core_workload_insertion_retry_interval = 3

# Distributed Tracing via Apache HTrace (http://htrace.incubator.apache.org/)
#
# Defaults to blank / no tracing
# Below sends to a local file, sampling at 0.1%
#
# htrace.sampler.classes=ProbabilitySampler
# htrace.sampler.fraction=0.001
# htrace.span.receiver.classes=org.apache.htrace.core.LocalFileSpanReceiver
# htrace.local.file.span.receiver.path=/some/path/to/local/file
#
# To capture all spans, use the AlwaysSampler
#
# htrace.sampler.classes=AlwaysSampler
#
# To send spans to an HTraced receiver, use the below and ensure
# your classpath contains the htrace-htraced jar (i.e. when invoking the ycsb
# command add -cp /path/to/htrace-htraced.jar)
#
# htrace.span.receiver.classes=org.apache.htrace.impl.HTracedSpanReceiver
# htrace.htraced.receiver.address=example.com:9075
# htrace.htraced.error.log.period.ms=10000