Skip to main content
Vulnerability Database/CVE-2026-54426

CVE-2026-54426: Apache bRPC Redis Client DOS Vulnerability

CVE-2026-54426 is a denial of service flaw in Apache bRPC Redis client that allows remote servers to crash client processes through deeply nested replies. 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-2026-54426 Overview

CVE-2026-54426 is an uncontrolled recursion vulnerability in the Redis reply parser of Apache bRPC versions prior to 1.18.0. A remote Redis server peer can crash a bRPC Redis client process by returning a deeply nested RESP (REdis Serialization Protocol) array reply. The parser recurses once for each nested array level without enforcing any depth limit. This causes stack exhaustion and a denial-of-service condition in the client process. The flaw is tracked under CWE-674: Uncontrolled Recursion.

Critical Impact

A malicious or compromised Redis server can crash any bRPC-based client configured with ChannelOptions.protocol="redis" by sending a crafted deeply nested RESP array reply.

Affected Products

  • Apache bRPC versions prior to 1.18.0
  • Applications using bRPC as a Redis client with ChannelOptions.protocol="redis"
  • Services receiving replies from untrusted or faulty Redis server peers

Discovery Timeline

  • 2026-10-21 - CVE-2026-54426 published to NVD
  • 2026-10-22 - Last updated in NVD database

Technical Details for CVE-2026-54426

Vulnerability Analysis

Apache bRPC implements a Redis client that parses responses conforming to the RESP protocol. RESP supports an array type that can contain nested arrays as elements. The bRPC reply parser handles nested arrays by making a recursive function call for each additional nesting level. The parser does not enforce a maximum recursion depth before invoking itself again.

A remote Redis server peer can craft a reply containing hundreds or thousands of nested array markers. When the bRPC client processes this response, each level of nesting consumes a stack frame. The process exhausts its thread stack and terminates with a segmentation fault. The result is a denial-of-service against the client application, not the Redis server.

Root Cause

The root cause is the absence of a depth limit in the Redis reply parsing routine. The parser trusts the structure of the inbound reply and recurses unconditionally. This pattern maps directly to CWE-674: Uncontrolled Recursion. The fix introduced in PR #3344 adds a configurable redis_max_reply_depth option with a default value of 128. Replies exceeding that depth are rejected before further recursion occurs.

Attack Vector

Exploitation requires that a bRPC client connect to a Redis endpoint controlled or influenced by an attacker. Scenarios include a compromised Redis server, a man-in-the-middle on an unencrypted Redis connection, or a shared cache populated with attacker-supplied values returned as nested arrays. The attacker returns a RESP reply with array nesting that exceeds the client's available stack depth. The bRPC client process crashes, interrupting any service that depends on it. See the Apache mailing list disclosure for the full advisory.

Detection Methods for CVE-2026-54426

Indicators of Compromise

  • Unexpected crashes or segmentation faults in processes linking Apache bRPC configured as a Redis client
  • Stack traces showing deep recursion inside the bRPC Redis reply parsing functions
  • Repeated client-side reconnects to a Redis endpoint followed by immediate termination

Detection Strategies

  • Inventory services that depend on Apache bRPC and identify those using ChannelOptions.protocol="redis" against versions earlier than 1.18.0
  • Inspect Redis traffic for RESP responses containing anomalously deep array nesting beyond typical application patterns
  • Correlate crash events in bRPC clients with network sessions to Redis endpoints that return malformed or oversized replies

Monitoring Recommendations

  • Track process crash and restart telemetry for bRPC-based services and alert on repeated failures tied to Redis connectivity
  • Collect core dumps from affected clients and review for recursion in the RESP parser call stack
  • Monitor Redis server integrity and authentication logs to identify unauthorized responses or compromised endpoints

How to Mitigate CVE-2026-54426

Immediate Actions Required

  • Upgrade Apache bRPC to version 1.18.0 or later across all affected services
  • Audit all bRPC Redis client connections to confirm they terminate only at trusted, authenticated Redis endpoints
  • Enforce TLS and authentication on Redis channels to prevent injection of malicious replies by on-path attackers

Patch Information

Apache bRPC 1.18.0 contains the fix. The patch, available in Apache bRPC PR #3344, introduces a configurable redis_max_reply_depth option with a default value of 128. Replies with array nesting deeper than this limit are rejected before recursion continues. Operators requiring deeper legitimate nesting can tune the value, while those with stricter requirements can lower it.

Workarounds

  • Restrict bRPC Redis clients to connect only to Redis servers under direct administrative control
  • Place Redis endpoints behind network segmentation and mutual authentication to prevent untrusted peers from returning crafted replies
  • If upgrading is not immediately feasible, apply the patch from PR #3344 directly to the deployed bRPC source tree and rebuild
bash
# Configuration example: tune the new depth limit after upgrading to bRPC 1.18.0
# Lower the default (128) if your workload never returns deeply nested RESP arrays
--redis_max_reply_depth=32

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.