MapReduce section

The MapReduce section of the application profile contains settings related to the MapReduce application and resources assigned to the MapReduce consumer associated with the application. This section is mandatory for MapReduce applications and must be defined only once in each application profile.

Note: This section is not supported for IBM® Spectrum Symphony applications. If your IBM Spectrum Symphony application defines this section, ensure that you remove the section and re-register the application; otherwise, the application will be rejected.

MapReduce

Description: Defines MapReduce settings.

MapReduce > ResourceGroups

Description: Resource groups from which resources are requested to run service instance managers and services to perform computations for clients. The resource group must contain compute hosts, not management hosts.
reduceLimitPerHost
Description: Specifies the maximum number of reduce tasks that can run on a host.

Where used: MapReduce > ResourceGroups

Required or optional: Optional

Valid values: 0 to 2147483647. A value of -1 indicates an unlimited number of reduce tasks can be run on a host.

Default value: -1

runMapTaskOnReduceSlot
Description: Specifies whether map tasks can run on reduce slots, when the number of available map slots is less than those required to run map tasks but the number of available reduce slots exceeds those required to run reduce tasks.

Where used: MapReduce > ResourceGroups

Required or optional: Optional

Valid values: true or false

Default value: true

runReduceTaskOnMapSlot
Description: Specifies whether reduce tasks can run on map slots, when the number of available reduce slots is less than those required to run reduce tasks but the number of available map slots exceeds those required to run map tasks.

Where used: MapReduce > ResourceGroups

Required or optional: Optional

Valid values: true or false

Default value: true

targetMrRatio
Description: Specifies the target slot ratio for running map tasks and reduce tasks in the resource group. Based on this map-to-reduce ratio, the SSM (application manager) divides resources into two pools: the map pool and the reducer pool. When the SSM is allocated multiple slots on a host, the policy tries to balance the M:R ratio on each host. When the SSM is allocated a single slot on a host, the policy tries to balance the overall M:R ratio for the policy.

Where used: MapReduce > ResourceGroups

Required or optional: Optional

Valid values: X:Y, where both X and Y are positive integers with a maximum of 100 and at least one of them is not zero.

Default value: 1:1

Examples: 0:2, 2:0, or 3:4

ResourceGroups > ResourceGroup

Description: Defines the configuration for each resource group.
name
Description: Resource group from which resources are requested to run service instance managers and services to perform computations for clients. The resource group name must be unique within the cluster and must contain compute hosts, not management hosts. The resource group name cannot contain multiple resource groups (for example, rg1 rg2 rg3), but you can use the asterisk (*) character to specify a common configuration for all resource groups associated with the MapReduce consumer (MapReduceConsumer). Each resource group must be defined only once, with the hosts in each resource group belonging only to that resource group.

Not required for IBM Spectrum Symphony Developer Edition.

Where used: MapReduce > ResourceGroups > ResourceGroup

Required or optional: Required

Valid values: *, or any alphanumeric string that does not include any special characters:
  • * indicates all resource groups available to the MapReduce consumer (MapReduceConsumer). For example, name="*".
  • Any alphanumeric string that does not include any special characters will override settings specified for resource groups included in *.
reduceLimitPerHost
Description: Specifies the maximum number of reduce tasks that can run on a host.

If you specify this setting, this value overrides the value specified in ResourceGroups. If this setting is not specified, the value specified in ResourceGroups will apply. If ResourceGroup name is specified as * (that is, name="*"), this setting applies to all resource groups associated with the consumer (MapReduceConsumer).

Where used: MapReduce > ResourceGroups > ResourceGroup

Required or optional: Optional

Valid values: 0 or more. A value of -1 indicates an unlimited number of reduce tasks can be run on a host.

Default value: -1

runMapTaskOnReduceSlot
Description: Specifies whether map tasks can run on reduce slots, when the number of available map slots is less than those required to run map tasks but the number of available reduce slots exceeds those required to run reduce tasks.

If you specify this setting, this value overrides the value specified in ResourceGroups. If this setting is not specified, the value specified in ResourceGroups will apply. If ResourceGroup name is specified as * (that is, name="*"), this setting applies to all resource groups associated with the consumer (MapReduceConsumer).

Where used: MapReduce > ResourceGroups > ResourceGroup

Required or optional: Optional

Valid values: true or false

Default value: true

runReduceTaskOnMapSlot
Description: Specifies whether reduce tasks can run on map slots, when the number of available reduce slots is less than those required to run reduce tasks but the number of available map slots exceeds those required to run map tasks.

If you specify this setting, this value overrides the value specified in ResourceGroups. If this setting is not specified, the value specified in ResourceGroups will apply. If ResourceGroup name is specified as * (that is, name="*"), this setting applies to all resource groups associated with the consumer (MapReduceConsumer).

Where used: MapReduce > ResourceGroups > ResourceGroup

Required or optional: Optional

Valid values: true or false

Default value: true

targetMrRatio
Description: Specifies the target slot ratio for running map tasks and reduce tasks in the specified resource group. Based on this map-to-reduce ratio, the SSM (application manager) divides resources into two pools: the map pool and the reducer pool. When the SSM is allocated multiple slots on a host, the policy tries to balance the M:R ratio on each host. When the SSM is allocated a single slot on a host, the policy tries to balance the overall M:R ratio for the policy.

If you specify this setting, this value overrides the value specified in ResourceGroups. If this setting is not specified, the setting specified in ResourceGroups will apply. If ResourceGroup name is specified as * (that is, name="*"), this setting applies to all resource groups associated with the consumer (MapReduceConsumer).

Where used: MapReduce > ResourceGroups > ResourceGroup

Required or optional: Optional

Valid values: X:Y, where both X and Y are positive integers with a maximum of 100 and at least one of them is not zero.

Default value: 1:1

Examples: 0:2, 2:0, or 3:4