Skip to main content
Vulnerability Database/CVE-2026-82425

CVE-2026-82425: Apache bRPC Denial of Service Vulnerability

CVE-2026-82425 is a denial of service flaw in Apache bRPC that allows remote attackers to crash processes through assertion failures. This post explains its technical details, 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-2026-82425 Overview

CVE-2026-82425 is a reachable assertion vulnerability [CWE-617] in Apache bRPC versions prior to 1.18.0. The flaw exists in the Socket lifecycle handling when health checking is enabled. A malicious remote peer can race connection teardown against health check revival to trigger a CHECK assertion inside Socket::OnInputEvent. When bRPC is built with glog, with -crash_on_fatal_log enabled, or in OFFICIAL_BUILD/NDEBUG configurations, the assertion terminates the process. The result is a reliable denial of service against bRPC clients that connect to attacker-controlled endpoints.

Critical Impact

A remote peer can abort client processes using Apache bRPC health checking, causing service outages across production RPC fleets.

Affected Products

  • Apache bRPC versions prior to 1.18.0
  • Applications using bRPC clients with -health_check_interval enabled
  • Builds using glog, -crash_on_fatal_log, OFFICIAL_BUILD, or NDEBUG

Discovery Timeline

  • 2026-10-21 - CVE-2026-82425 published to NVD
  • 2026-10-22 - Last updated in NVD database

Technical Details for CVE-2026-82425

Vulnerability Analysis

The vulnerability resides in the interaction between bRPC's Socket object lifecycle and its event dispatcher. When a Socket with health checking enabled fails, Socket::WaitAndReset closes the underlying file descriptor. The event dispatcher may still deliver an input event (EPOLLIN on Linux or EVFILT_READ on BSD/macOS) that was fetched for the now-closed file descriptor.

If the health check probe succeeds and revives the Socket, Socket::Address returns successfully even while the file descriptor remains -1. The next write operation reconnects the socket on demand. Inside this race window, the stale input event is routed to Socket::OnInputEvent, which invokes CHECK(!(events & EPOLLIN)). The assertion fails because the Socket is addressable but has no file descriptor, terminating the process.

Root Cause

The root cause is a time-of-check to time-of-use condition between the event dispatcher and the health check revival path. The code assumes that an addressable Socket always has a valid file descriptor when input events arrive. The health check revival path breaks that invariant without draining or filtering stale events queued for the previous file descriptor.

Attack Vector

A malicious remote peer that a bRPC client connects to can repeatedly abort TCP connections and immediately accept subsequent health check probes. By tuning the timing of connection teardown and probe acceptance, the attacker races the event dispatcher against health check revival. This triggers the assertion failure and aborts the client process. Exploitation requires only that the client has -health_check_interval enabled and connects to the attacker-controlled endpoint.

See the Apache Mailing List Thread for the full technical description and the upstream fix in pull request #3510.

Detection Methods for CVE-2026-82425

Indicators of Compromise

  • Unexpected process termination of bRPC client applications with CHECK failure messages in logs referencing Socket::OnInputEvent or EPOLLIN.
  • Fatal glog entries containing crash_on_fatal_log or assertion traces originating from brpc::Socket internals.
  • Repeated short-lived TCP connections followed by immediate reconnects from the same remote peer prior to a crash.

Detection Strategies

  • Inventory bRPC-linked binaries and verify version strings against 1.18.0 to identify vulnerable builds.
  • Correlate process crash telemetry with outbound RPC connection logs to identify peers that preceded the termination.
  • Alert on repeated client restarts tied to specific remote endpoints, which indicate iterative exploitation attempts.

Monitoring Recommendations

  • Forward crash and core dump telemetry from bRPC client hosts into a central analytics pipeline for pattern analysis.
  • Monitor connection churn metrics such as reconnect rate and health check failure rate on bRPC clients.
  • Track glog fatal log counters and alert on sudden increases, which may indicate exploitation of reachable assertions.

How to Mitigate CVE-2026-82425

Immediate Actions Required

  • Upgrade Apache bRPC to version 1.18.0 or later across all client deployments.
  • Audit application configurations for the -health_check_interval flag and identify services exposed to untrusted peers.
  • Restrict bRPC client connectivity to trusted upstream endpoints until patching is complete.

Patch Information

The Apache bRPC project fixed this issue in version 1.18.0. The upstream patch is tracked in Apache bRPC pull request #3510. After applying the fix, stale input events on a Socket revived without a file descriptor are silently dropped instead of triggering the assertion. The fix changes no user-visible interfaces and no wire protocol behavior.

Workarounds

  • Disable health checking by removing -health_check_interval where operationally acceptable until the patch is applied.
  • Rebuild bRPC without -crash_on_fatal_log and avoid OFFICIAL_BUILD/NDEBUG configurations if downtime from assertion failures is unacceptable.
  • Place bRPC clients behind trusted proxies or allowlists so that only vetted peers can respond to health check probes.

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.