AWS emr documentation change
Summary
Corrected a typo in the configuration property name from 'yarn.resourcemager.decommissioning.timeout' to 'yarn.resourcemanager.decommissioning.timeout' in the Spark node decommissioning documentation.
Security assessment
This change fixes a documentation typo in a configuration property name but does not address any security vulnerability, weakness, or incident. The documentation describes Spark's node decommissioning behavior for reliability during Spot instance terminations, not security controls.
Diff
diff --git a/emr/latest/ReleaseGuide/emr-spark-configure.md b/emr/latest/ReleaseGuide/emr-spark-configure.md index 8f0c4b0c2..6c36ff8ac 100644 --- a//emr/latest/ReleaseGuide/emr-spark-configure.md +++ b//emr/latest/ReleaseGuide/emr-spark-configure.md @@ -121 +121 @@ Spark node decommissioning settings Setting | Description | Default value -`spark.decommissioning.timeout.threshold` | Available in Amazon EMR release 5.11.0 or later. Specified in seconds. When a node transitions to the decommissioning state, if the host will decommission within a time period equal to or less than this value, Amazon EMR not only deny lists the node, but also cleans up the host state (as specified by `spark.resourceManager.cleanupExpiredHost`) without waiting for the node to transition to a decommissioned state. This allows Spark to handle Spot instance terminations better because Spot instances decommission within a 20-second timeout regardless of the value of `yarn.resourcemager.decommissioning.timeout`, which may not provide other nodes enough time to read shuffle files. | `20s` +`spark.decommissioning.timeout.threshold` | Available in Amazon EMR release 5.11.0 or later. Specified in seconds. When a node transitions to the decommissioning state, if the host will decommission within a time period equal to or less than this value, Amazon EMR not only deny lists the node, but also cleans up the host state (as specified by `spark.resourceManager.cleanupExpiredHost`) without waiting for the node to transition to a decommissioned state. This allows Spark to handle Spot instance terminations better because Spot instances decommission within a 20-second timeout regardless of the value of `yarn.resourcemanager.decommissioning.timeout`, which may not provide other nodes enough time to read shuffle files. | `20s`