YCSB/crail

Quick Start

This section describes how to run YCSB on Apache Crail.

1. Start Crail

https://incubator-crail.readthedocs.io

2. Install Java and Maven

3. Set Up YCSB

Git clone YCSB and compile:

git clone http://github.com/brianfrankcooper/YCSB.git
cd YCSB
mvn -pl com.yahoo.ycsb:crail-binding -am clean package

4. Provide Crail Connection Parameters

Connction parameters have to be defined in $CRAIL_HOME/conf/crail-site.conf.

• crail.namenode.address - The Crail cluster to connect to (default: crail://namenode:9060)
• crail.blocksize - The block size (bytes) of the Crail cluster (default: 1048576)
• crail.buffersize - The buffer size (bytes) used by the client (default: crail.blocksize)
• crail.cachelimit - Maximum client side cache (bytes) (default: 1073741824)
• crail.cachepath - Directory where to mmap memory from (no default)
• crail.storage.types - Comma separated list of storage tiers (default: org.apache.crail.storage.tcp.TcpStorageTier)

The following benchmark parameters are available.

• crail.enumeratekeys - Whether to make keys visible for enumeration or not (default: false)

Add them to the workload or set them with the shell command, as in:

./bin/ycsb load crail -s -P workloads/workloada -p crail.enumeratekeys=true >outputLoad.txt

5. Load Data and Run Tests

Load the data:

./bin/ycsb load crail -s -P workloads/workloada 

Run the workload test:

./bin/ycsb run crail -s -P workloads/workloada