[ad_1]
Researchers have revealed particulars of a now-patched high-severity safety vulnerability in Apache Cassandra that, if left unaddressed, may very well be abused to realize distant code execution on affected installations.
“This Apache safety vulnerability is straightforward to take advantage of and has the potential to wreak havoc on programs, however fortunately solely manifests in non-default configurations of Cassandra,” Omer Kaspi, safety researcher at DevOps agency JFrog, mentioned in a technical write-up printed Tuesday.
Apache Cassandra is an open-source, distributed, NoSQL database administration system for managing very massive quantities of structured knowledge throughout commodity servers.
Tracked as CVE-2021-44521 (CVSS rating: 8.4), the vulnerability considerations a selected state of affairs the place the configuration for user-defined capabilities (UDFs) are enabled, successfully permitting an attacker to leverage the Nashorn JavaScript engine, escape the sandbox, and obtain execution of untrusted code.
Particularly, it was discovered that Cassandra deployments are susceptible to CVE-2021-44521 when the cassandra.yaml configuration file accommodates the next definitions:
- enable_user_defined_functions: true
- enable_scripted_user_defined_functions: true
- enable_user_defined_functions_threads: false
“When the [enable_user_defined_functions_threads] choice is ready to false, all invoked UDF capabilities run within the Cassandra daemon thread, which has a safety supervisor with some permissions,” Kaspi mentioned, thereby permitting the adversary to disable the safety supervisor and get away of the sandbox and run arbitrary shell instructions on the server.
Apache Cassandra customers are inspired to improve to variations 3.0.26, 3.11.12, and 4.0.2 to keep away from potential exploitation, which addresses the flaw by including a brand new flag “allow_extra_insecure_udfs” that is set to false by default and prevents turning off the safety supervisor.
[ad_2]
