Databricks Mitigates Java4Shell Vulnerability Before Public Disclosure

Databricks says it had already begun addressing the industry-wide Java4Shell OpenJDK vulnerability, CVE-2026-114938, before its public disclosure. The company has disabled use of the vulnerable code across supported Databricks Runtime versions and will follow with patched releases.
- Databricks addressed Java4Shell, CVE-2026-114938, before public disclosure.
- The mitigation disables vulnerable code across supported Databricks Runtime versions.
- Affected clusters were proactively restarted to apply the mitigation.
- Serverless customers will receive patched runtimes automatically.
- Databricks reported no evidence of unauthorized access to products, services, or customer data.
Mitigation applied across supported runtimes
In an update dated November 19 at 10 AM Pacific Time, Databricks said it is protecting customers in two phases. The first phase is a mitigation designed to disable use of the vulnerable code, while the second will replace that code with patched versions.
According to Databricks, the mitigation has been applied to its platform and disables use of the vulnerable code for all supported versions of Databricks Runtime. The company said it does not expect the measure to have a meaningful performance impact on customers.
Affected clusters were proactively restarted
Databricks said it took the rare step of proactively restarting affected clusters so it could apply the mitigation as quickly as possible. The company described the mitigation as a measure that blocks the attack while it completes the next stage of its response.
Databricks will provide patched versions of supported Databricks Runtimes. This patching phase is intended to replace the vulnerable code, following the immediate mitigation that disables its use.
What customers need to do
Customers using the Databricks serverless platform will receive the patched runtimes automatically, according to the company. Customers with classic clusters will need to restart those clusters.
Databricks also said it is reaching out to customers running custom images. The announcement does not provide additional details about those images or specify a date for the patched runtime versions.
Databricks reports no evidence of unauthorized access
Databricks said it had no evidence of unauthorized access to its products, services, or customers' data. The company's two-phase response means the vulnerable code has been disabled on supported Databricks Runtime versions, while patched versions are being provided as a further precaution.
For users, the immediate actions depend on their deployment type. Serverless customers are set to receive the patched versions automatically, while customers operating classic clusters must restart them to obtain the updates.
Databricks: Java4Shell vulnerability (CVE-2026-114938) assessment


