[SPARK-56158][CORE] Support limitActiveProcessorCount in local mode#55132
Open
jzhuge wants to merge 1 commit intoapache:masterfrom
Open
[SPARK-56158][CORE] Support limitActiveProcessorCount in local mode#55132jzhuge wants to merge 1 commit intoapache:masterfrom
jzhuge wants to merge 1 commit intoapache:masterfrom
Conversation
In local mode, the driver runs in the same JVM as spark-submit. When spark.driver.limitActiveProcessorCount.enabled=true, inject -XX:ActiveProcessorCount=<spark.driver.cores> in SparkSubmitCommandBuilder, unless the user already set it in driver Java options.
Member
Author
|
Good to go |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Extend
spark.driver.limitActiveProcessorCount.enabledto local mode. Previously the flag only worked in YARN cluster mode (yarn/Client.scala).In local mode the driver runs in the same JVM as
spark-submit, so the JVM flag must be injected before the JVM starts. The injection is added inSparkSubmitCommandBuilder, which builds the JVM command line prior to exec:-XX:ActiveProcessorCount=<spark.driver.cores>when the master islocalorlocal[...](local-clusteris excluded — it spawns separate worker processes)-XX:ActiveProcessorCountin driver Java options--driver-coresinOptionParser(consistent with--driver-memory) so it is available in the effective configWhy are the changes needed?
spark.driver.limitActiveProcessorCount.enabledhad no effect in local mode. Users on multi-core machines couldn't limit driver CPU with the same mechanism available in YARN.Does this PR introduce any user-facing change?
Yes.
spark.driver.limitActiveProcessorCount.enabled=truenow takes effect in local mode:-XX:ActiveProcessorCount=<spark.driver.cores>(default 1) is injected into the JVM command byspark-submit.How was this patch tested?
Added
testLimitActiveProcessorCountLocalModeinSparkSubmitCommandBuilderSuitecovering: flag disabled, enabled with default/custom cores (--confand--driver-cores), YARN client mode (no injection),local-cluster(no injection), user-supplied flag (no duplication), invalidspark.driver.cores(throws).Was this patch authored or co-authored using generative AI tooling?
Yes