Skip to main content
Vulnerability Database/CVE-2026-114236

CVE-2026-114236: Storage Archive Layer DOS Vulnerability

CVE-2026-114236 is a denial of service flaw in storage archive layer processing that causes file descriptor exhaustion when handling malicious inputs. This post explains its impact, affected versions, and mitigation steps.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-114236 Overview

CVE-2026-114236 is a file descriptor leak vulnerability in the storage/pkg/archive Go package, specifically affecting the ApplyLayer and ApplyUncompressedLayer functions. When these functions process corrupt or malicious archive inputs, they fail to properly release file descriptors. Repeated, unbounded invocations against malformed layers can exhaust the host process's file descriptor limit, leading to a local denial of service condition. The flaw is categorized under CWE-772: Missing Release of Resource after Effective Lifetime. The package is commonly used by container runtimes and image-handling tooling that process OCI and Docker image layers.

Critical Impact

Attackers with the ability to supply crafted container image layers can exhaust process file descriptors, degrading or halting container image operations.

Affected Products

  • Projects embedding the containers/storage Go module storage/pkg/archive package
  • Container runtime and image tooling consuming ApplyLayer and ApplyUncompressedLayer
  • Downstream distributions tracked via the Red Hat CVE-2026-114236 Advisory

Discovery Timeline

  • 2026-10-21 - CVE-2026-114236 published to NVD
  • 2026-10-22 - Last updated in NVD database

Technical Details for CVE-2026-114236

Vulnerability Analysis

The ApplyLayer and ApplyUncompressedLayer functions in storage/pkg/archive extract tar-based filesystem layers onto a target directory. During extraction, the functions open file descriptors for entries within the archive. When the input stream is malformed or contains deliberately corrupted entries, error-handling paths return without closing descriptors that were already opened for in-progress entries. Each failed invocation leaks a descriptor. Because the leaked descriptors persist for the lifetime of the calling process, repeated calls against malicious inputs steadily consume the process's RLIMIT_NOFILE budget. Once the limit is reached, subsequent file, socket, and pipe operations fail, which disrupts container pulls, builds, and other runtime activity dependent on the archive package.

Root Cause

The root cause is a missing release of file descriptor resources on error paths within the layer-application logic. The functions lack a guaranteed cleanup mechanism, such as a defer-based close, that executes when parsing or extraction fails partway through an entry. The condition aligns with [CWE-772].

Attack Vector

Exploitation requires local access with the ability to feed crafted archive layers to a process that invokes the affected functions. User interaction is involved because a victim or automation workflow must trigger extraction of the attacker-supplied layer. No privileges are required to supply the input, and the impact is limited to availability of the targeted process. See the Red Hat Bug Report #2558105 for additional context.

Detection Methods for CVE-2026-114236

Indicators of Compromise

  • Rising open file descriptor counts for container runtime processes without a corresponding increase in legitimate workload activity.
  • Repeated EMFILE or too many open files errors emitted by processes using containers/storage.
  • Container image pull or layer extraction failures correlated with specific untrusted image sources.

Detection Strategies

  • Monitor /proc/<pid>/fd counts for long-lived processes that call into storage/pkg/archive and alert on sustained upward trends.
  • Inspect runtime logs for errors returned by ApplyLayer and ApplyUncompressedLayer paired with resource-exhaustion signals.
  • Audit image registries and build pipelines for layer artifacts that fail decompression or tar parsing.

Monitoring Recommendations

  • Track per-process file descriptor usage against configured RLIMIT_NOFILE and alert at defined thresholds.
  • Collect and centralize container runtime and build logs so repeated extraction failures from a single source are visible.
  • Baseline normal image-pull failure rates and investigate deviations tied to untrusted registries.

How to Mitigate CVE-2026-114236

Immediate Actions Required

  • Inventory services and tooling that embed containers/storage and identify versions in use.
  • Restrict ingestion of container images to trusted registries until patched builds are deployed.
  • Raise RLIMIT_NOFILE for affected processes as a short-term buffer while patches are rolled out.

Patch Information

Refer to the Red Hat CVE-2026-114236 Advisory for distribution-specific fixed package versions. Upstream fixes address the error-path cleanup in ApplyLayer and ApplyUncompressedLayer so file descriptors are released regardless of parse outcome. Rebuild and redeploy any statically linked Go binaries that vendor the affected module.

Workarounds

  • Reject or quarantine container images from untrusted sources until the patched module is in place.
  • Restart long-running processes that invoke the affected functions on a schedule to reclaim leaked descriptors.
  • Apply process-level resource limits and automatic restart policies to contain the availability impact.

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.