The behavior of Tez runtime is specified by the configuration file tez-site.xml in the classpath. MR3 inherits all configuration keys for Tez runtime from original Tez. For example, specifies the amount of memory required for sorting the output.

MR3 also introduces additional configuration keys which are specific to new features of MR3, and may interpret existing configuration keys in a different way. Below we describe such configuration keys.


Name Default value Description
tez.runtime.pipelined.sorter.use.soft.reference false true: use soft references for ByteBuffers allocated in PipelinedSorter. These soft references are reused across TaskAttempts running in the same ContainerWorker. false: do not use soft references. For more details, see Basic Performance Tuning. false true: enable auto parallelism for ShuffleVertexManager. false: disable auto parallelism. For more details, see Auto Parallelism and Basic Performance Tuning. 20 Minimum number of Tasks to trigger auto parallelism. For example, if the value is set to 20, only those Vertexes with at least 20 Tasks are considered for auto parallelism. The user can effectively disable auto parallelism by setting this configuration key to a large value. 10 Specifies the percentage of Tasks that can be kept after applying auto parallelism. For example, if the value is set to 10, the number of Tasks can be reduced by up to 100 - 10 = 90 percent, thereby leaving 10 percent of Tasks.
tez.shuffle-vertex-manager.use-stats-auto-parallelism false true: analyze input statistics when applying auto parallelism. false: do not use input statistics. 20 Specifies the lower limit when normalizing input statistics. For example, if the value is set to 20, input statistics are normalized between 20 and 100. That is, an input size of zero is normalized to 20 while the maximum input size is mapped to 100.
Name Default value Description mapreduce_shuffle Service ID for the external shuffle service. Set to tez_shuffle in order to use the shuffle handler of MR3.
tez.shuffle.max.threads 0 Number of threads in each shuffle handler. With the default value of zero, each shuffle handler creates twice as many threads as the number of cores. For more details, see Basic Performance Tuning.
tez.shuffle.port 13563 Default port number for the shuffle handler of MR3
tez.runtime.shuffle.connect.timeout 12500 Maximum time in milliseconds for trying to connect to the shuffle service or the built-in shuffle handler before reporting fetch-failures. For more details, see Fault Tolerance and Basic Performance Tuning.
tez.shuffle.indexcache.mb 10 Size of path index cache in MB for MR3 shuffle handlers.
tez.shuffle.indexcache.share true true: All MR3 shuffle handlers share path index cache. false: Each MR3 shuffle handler uses its own path index cache. Set to false only when the number of shuffle handlers is very large. false true: If the size of free memory exceeds the size of memory allocated to a single Task, fetchers use MemoryFetchedInput (for unordered data) and InMemoryMapOutput and (for ordered data) instead of spilling to local disks. false: Fetchers do not consider the size of free memory. For more details, see Basic Performance Tuning
tez.runtime.shuffle.parallel.copies 20 Maximum number of fetchers per LogicalInput. Note that a RuntimeTask can create several LogicalInputs. 40 Maximum number of fetchers per ContainerWorker. 20 Maximum number of task output files to fetch per fetch request. A large value can produces HTTP 400 errors.
tez.runtime.shuffle.max.input.hostports 10000 Maximum number of host-port combinations to cache for shuffling (to prevent memory-leak in public clouds with autoscaling)
tez.runtime.shuffle.ranges.scheme first first: ShuffleServer selects randomly LogicalInput for shuffling. max (experimental): ShuffleServer selects LogicalInput with the most number of pending inputs.


The following configuration keys are effective when tez.celeborn.enabled is set to true and MR3 uses Celeborn as remote shuffle service. A configuration key of the form tez.celeborn.XXX.YYY is automatically converted to celeborn.XXX.YYY and passed to the Celeborn client.

Name Default value Description
tez.runtime.celeborn.fetch.split.threshold 1073741824 Maximum size of data (in bytes) that a fetcher can receive from Celeborn workers. The default value is 1GB.
tez.runtime.celeborn.shuffle.parallel.copies 4 Maximum number of fetchers that can be simultaneously active in a reducer.