This section describes how to run YCSB on Solr running locally.

Clone the YCSB git repository and compile:

git clone git:https://github.com/brianfrankcooper/YCSB.git
cd YCSB
mvn -pl site.ycsb:solr7-binding -am clean package

There must be a running Solr instance with a core/collection pre-defined and configured.

• See this API reference on how to create a core.
• See this API reference on how to create a collection in SolrCloud mode.

The conf/schema.xml configuration file present in the core/collection just created must be configured to handle the expected field names during benchmarking. Below illustrates a sample from a schema config file that matches the default field names used by the ycsb client:

<field name="id" type="string" indexed="true" stored="true" required="true" multiValued="false"/>
<field name="field0" type="text_general" indexed="true" stored="true"/>
<field name="field1" type="text_general" indexed="true" stored="true"/>
<field name="field2" type="text_general" indexed="true" stored="true"/>
<field name="field3" type="text_general" indexed="true" stored="true"/>
<field name="field4" type="text_general" indexed="true" stored="true"/>
<field name="field5" type="text_general" indexed="true" stored="true"/>
<field name="field6" type="text_general" indexed="true" stored="true"/>
<field name="field7" type="text_general" indexed="true" stored="true"/>
<field name="field8" type="text_general" indexed="true" stored="true"/>
<field name="field9" type="text_general" indexed="true" stored="true"/>

If running in SolrCloud mode ensure there is an external Zookeeper cluster running.

• See here for details on how to set up an external Zookeeper cluster.
• See here for instructions on how to use Zookeeper to manage your core/collection configuration files.

Now you are ready to run! First, load the data:

./bin/ycsb load solr7 -s -P workloads/workloada -p table=<core/collection name>

Then, run the workload:

./bin/ycsb run solr7 -s -P workloads/workloada -p table=<core/collection name>

For further configuration see below:

The default settings for the Solr node that is created is as follows:

• solr.cloud

  • A Boolean value indicating if Solr is running in SolrCloud mode. If so there must be an external Zookeeper cluster running also.
  • Default value is false and therefore expects solr to be running in stand-alone mode.

• solr.base.url

  • The base URL in which to interface with a running Solr instance in stand-alone mode
  • Default value is `https://localhost:8983/solr

• solr.commit.within.time

  • The max time in ms to wait for a commit when in batch mode, ignored otherwise
  • Default value is 1000ms

• solr.batch.mode

  • Indicates if inserts/updates/deletes should be commited in batches (frequency controlled by the solr.commit.within.time parameter) or commit 1 document at a time.
  • Default value is false

• solr.zookeeper.hosts

  • A list of comma seperated host:port pairs of Zookeeper nodes used to manage SolrCloud configurations.
  • Must be passed when in SolrCloud mode.
  • Default value is localhost:2181

If you wish to customize the settings used to create the Solr node you can created a new property file that contains your desired Solr node settings and pass it in via the parameter to 'bin/ycsb' script. Note that the default properties will be kept if you don't explicitly overwrite them.

Assuming that we have a properties file named "myproperties.data" that contains custom Solr node configuration you can execute the following to pass it into the Solr client:

./bin/ycsb run solr7 -P workloads/workloada -P myproperties.data -s

If you wish to use SolrCloud mode ensure a Solr cluster is running with an external zookeeper cluster and an appropriate collection has been created. Make sure to pass the following properties as parameters to 'bin/ycsb' script.

solr.cloud=true
solr.zookeeper.hosts=<zkHost2>:<zkPort1>,...,<zkHostN>:<zkPortN>