Provider Configuration Variables

When you configure an HDFS provider, Hunk automatically sets a number of configuration variables. You can use the preset variables, or you can modify them as needed by editing the provider.

• For more information about editing them in the configuration file, see "Set up a provider and virtual index in the configuration file."

• For information about editing providers in the Hunk user interface, see "Add an HDFS provider."

• For information about setting provider configuration variables for YARN, see "Required configuration variables for YARN."

Setting:	Use it to:
vix.splunk.setup.onsearch	Determines whether to perform setup (install & BR) on search.
vix.splunk.setup.package	Location for the Splunk .tgz package that Splunk can install and use on data nodes (in vix.splunk.home.datanode). A value of current uses current install.
vix.splunk.home.datanode	SPLUNK_HOME on the DataNode and/or TaskTracker
vix.splunk.home.hdfs	The location of scratch space on HDFS for this Splunk instance.
vix.splunk.search.debug	Determines whether search is run in debug mode.
vix.splunk.search.recordreader	Provides a comma separated list of data pre-processing classes This value must extend BaseSplunkRecordReader and return data to be consumed by Splunk as the value
vix.splunk.search.recordreader.avro.regex	Specifies a regex that files must match in order to be considered avro files, defaults to \.avro$.
vix.splunk.search.mr.threads	Determines the number of threads to use when reading map results from HDFS.
vix.splunk.search.mr.maxsplits	Determines the maximum number of splits in an MapReduce job.
vix.splunk.search.mr.poll	Determines the polling period for job status, in milliseconds.
vix.splunk.search.mixedmode	Determines whether mixed mode execution is enabled
vix.splunk.search.mixedmode.maxstream	Determines the maximum number of bytes to stream during mixed mode. The default value is 10GB. A value of 0 indicates that there is no stream limit. Bytes will cease streaming after the first split that takes the value over the limit.
vix.splunk.jars	Provides a comma-delimted list of dirs/jars to use in SH and MR

High-Availability configurations

HA-NN

Setting:	Use it to:
vix.fs.default.name	The name of the default file system uri> # hdfs://sveserv51-ha
vix.dfs.nameservices	Comma-separated list of nameservices> # sveserv51-ha
vix.dfs.ha.namenodes.sveserv51-ha	comma-separated list of namenodes for a given nameservice, eg sveserv51-ha> # nn1,nn2
vix.dfs.namenode.rpc-address.sveserv51-ha.nn1	The RPC server address and port for a given namenode, eg nn1, of a given nameservice, eg, sveserv51-ha> # sveserv51-vm6.sv.splunk.com:8020
vix.dfs.namenode.rpc-address.sveserv51-ha.nn2	The RPC server address and port for a given namenode, eg nn2, of a given nameservice, eg, sveserv51-ha> # sveserv51-vm5.sv.splunk.com:8020
vix.dfs.client.failover.proxy.provider.sveserv51-ha	A FailoverProxyProvider implementation for a given nameservice, eg, sveserv51-ha> # org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider

1. HA-JT

Setting:	Use it to:
vix.mapred.job.tracker	The logical name for a list of jobtrackers> # sveser51-ha-jt
vix.mapred.jobtrackers.sveserv51-ha-jt	comma-separated list of jobtrackers for a given logical jobtracker name, eg, sveserv51-ha-jt> # jt1,jt2
vix.mapred.jobtracker.rpc-address.sveserv51-ha-jt.jt1	The RPC server address and port for a given jobtracker, eg jt1, of a given logical jobtracker name, eg, sveserv51-ha-jt> # sveserv51-vm6.sv.splunk.com:8021
vix.mapred.jobtracker.rpc-address.sveserv51-ha-jt.jt2	The RPC server address and port for a given jobtracker, eg jt2, of a given logical jobtracker name, eg, sveserv51-ha-jt> # sveserv51-vm5.sv.splunk.com:8021
vix.mapred.client.failover. proxy.provider.sveserv51-ha-jt	A FailoverProxyProvider implementation for a given logical jobtracker name, eg, sveserv51-ha-jt> # org.apache.hadoop.mapred.ConfiguredFailoverProxyProvider
	vix.mapred.map.max.attempts = <Number of attempts a map task can be retired, until it's marked as failed> # 4
vix.mapred.max.map.failures.percent	Percent of acceptable task failures for the whole MR job, until the MR job fails> # 5