PT-2019-20133 · Apache · Apache Spark
Published
2019-08-07
·
Updated
2019-08-07
None
No severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
Name of the Vulnerable Software and Affected Versions
Apache Spark versions prior to 2.3.3
Description
The issue arises when Apache Spark writes user data to local disk unencrypted, despite having spark.io.encryption.enabled set to true. This occurs in specific situations, including when cached blocks are fetched to disk, controlled by spark.maxRemoteBlockSizeFetchToMem, and when using certain functions in SparkR and Pyspark, such as parallelize, broadcast, and python udfs.
Recommendations
For versions prior to 2.3.3, update to version 2.3.3 or later to resolve the issue. As a temporary workaround, consider disabling the use of spark.maxRemoteBlockSizeFetchToMem, parallelize, broadcast, and python udfs until a patch is available. Restrict access to sensitive data and minimize the use of affected functions to minimize the risk of exploitation.
Found an issue in the description? Have something to add? Feel free to write us 👾
Related Identifiers
Affected Products
Apache Spark