Three new security vulnerabilities have been discovered in Azure HDInsight’s Apache Hadoop, Kafka, and Spark services that could be exploited to achieve privilege escalation and a regular expression denial-of-service (ReDoS) condition.
“The new vulnerabilities affect any authenticated user of Azure HDInsight services such as Apache Ambari and Apache Oozie,” Orca security researcher Lidor Ben Shitrit said in a technical report shared with The Hacker News.
The list of flaws is as follows –
The two privilege escalation flaws could be exploited by an authenticated attacker with access to the target HDI cluster to send a specially crafted network request and gain cluster administrator privileges.
The XXE flaw is the result of a lack of user input validation that allows for root-level file reading and privilege escalation, while the JDBC injection flaw could be weaponized to obtain a reverse shell as root.
“The ReDoS vulnerability on Apache Oozie was caused by a lack of proper input validation and constraint enforcement, and allowed an attacker to request a large range of action IDs and cause an intensive loop operation, leading to a denial-of-service (DoS),” Ben Shitrit explained.
Successful exploitation of the ReDoS vulnerability could result in a disruption of the system’s operations, cause performance degradation, and negatively impact both the availability and reliability of the service.
Following responsible disclosure, Microsoft has rolled out fixes as part of updates released on October 26, 2023.
The development arrives nearly five months after Orca detailed a collection of eight flaws in the open-source analytics service that could be exploited for data access, session hijacking, and delivering malicious payloads.
In December 2023, Orca also highlighted a “potential abuse risk” impacting Google Cloud Dataproc clusters that take advantage of a lack of security controls in Apache Hadoop’s web interfaces and default settings when creating resources to access any data on the Apache Hadoop Distributed File System (HDFS) without any authentication.
