CVE-2020-9480

In Apache Spark 2.4.5 and earlier, a standalone resource manager's master may be configured to require authentication (spark.authenticate) via a shared secret

published June 23, 2020

CVSS
โ€”
EPSS
29%
Percentile
98.2
In the wild
Unconfirmed
What it is

โš ๏ธ A crafty RPC can bypass authentication in Apache Spark 2.4.5 and earlier, turning a trusted resource manager into an unwitting accomplice! ๐Ÿ”ฅ Think of it as a restaurant that requires a password for entry, but an intruder finds a way to place their order anyway โ€” the kitchen prepares the food without checking their credentials! An attacker could execute shell commands on the host machine, gaining control over critical resources and potentially wreaking havoc on your applications. It's a sneaky way to bypass security and could lead to unauthorized access and data breaches.

Put simply

Think of it as a restaurant that requires a password for entry, but an intruder finds a way to place their order anyway โ€” the kitchen prepares the food without checking their credentials! The vulnerability allows a specially-crafted remote procedure call (RPC) to the Spark master, which can start application resources without the necessary shared secret, effectively bypassing authentication.

What to do

An attacker could execute shell commands on the host machine, gaining control over critical resources and potentially wreaking havoc on your applications. It's a sneaky way to bypass security and could lead to unauthorized access and data breaches. To mitigate this vulnerability, upgrade Apache Spark to version 2.4.6 or later. Additionally, review your configuration to ensure that spark.authenticate is properly set and consider using more secure authentication methods. You've got this! By taking timely action, you'll be one step closer to securing your Spark cluster like a pro! ๐Ÿ›ก๏ธ

The record
Technical detail
CVSS
Not scored
CVSS v4.0
Not supplied
EPSS
0.29370 ยท 98.2th percentile
Published
2020-06-23T21:50Z
EPSS history
Timeline
  • 23 JUN 21:50Z
    In Apache Spark 2.4.5 and earlier, a standalone resource manager's master may be configured to require authentication (spark.authenticate) via a shared secret
    cvelistv5