Skip to content
VulniPulse
Advisory severityHigh7.5Red Hat Linux

High [CVE-2026-94627] Denial of Service via GPU memory exhaustion

This high-severity Red Hat Linux advisory covers CVE-2026-94627; related products: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI).

Aggregated and source-linked by VulniPulse. Data sources, validation and limitations.

CVE-2026-94627 Source published Source updated

VulniPulse record published Record updated

Related products & platforms
Red Hat LinuxUnclassified
Open source advisory

Android app · Google Play

Monitor future Red Hat Linux CVEs from your phone.

Choose a whole vendor or a precise platform, then receive matching security advisories by phone notification, email, or both. Coverage follows 32 official vendor sources and 160+ reviewed platform categories.

Matching phone alertsOptional email delivery

Summary

Denial of Service via GPU memory exhaustion. Red Hat rates this important (CVSS 7.5).

Weakness: CWE-772.

Affected products named by the advisory: Red Hat AI Inference Server; Red Hat Enterprise Linux AI (RHEL AI) 3; Red Hat OpenShift AI (RHOAI).

Affected versions
  • 0.29.0

Official advisory · high-confidence parse· fetched 8 days ago·verify at source

Fixed versions

No fixed release is recorded yet. That does not prove no patch exists — confirm against the vendor advisory.

Official advisory · high-confidence parse· fetched 8 days ago·verify at source

Mitigation checklist

Recommended fix / mitigation
  • Avoid using the Mooncake KV transfer connector in disaggregated prefill and decode topologies. Deploy inference workloads in standard standalone mode where prefill and decode are handled within the same instance. Additionally, restrict network access to the model serving endpoints to authenticated and authorized clients using Red Hat OpenShift NetworkPolicies or API gateway rate limiting. Caveats: Running without disaggregated prefill and decode execution can reduce throughput under heavy concurrent traffic. Network filtering requires callers to be within trusted cluster namespaces or behind authenticated proxies. Warning: Modifying serving runtime deployment manifests or restarting inference services will cause temporary service interruption while pods are recreated.

Official advisory · high-confidence parse· fetched 8 days ago·verify at source

Discussion(0)

No comments yet. Share field notes, upgrade gotchas, or questions — verify against the vendor advisory before acting on community advice.

Sign in to join the discussion.