Skip to main content
Vulnerability Database/CVE-2025-27100

CVE-2025-27100: lakeFS Memory Exhaustion DOS Vulnerability

CVE-2025-27100 is a memory exhaustion denial-of-service flaw in lakeFS that allows authenticated users to crash the server. 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-2025-27100 Overview

lakeFS is an open-source tool that transforms object storage into a Git-like repository. CVE-2025-27100 is an authenticated denial-of-service vulnerability affecting lakeFS versions 1.49.1 and below. An authenticated user can crash the lakeFS server by exhausting its memory. The issue is tracked as [CWE-400] Uncontrolled Resource Consumption and has been patched in version 1.50.0. The vulnerability is remotely exploitable over the network but requires valid authentication credentials.

Critical Impact

Any authenticated lakeFS user can trigger server memory exhaustion, crashing the service and disrupting access to data stored in the Git-like object storage repository.

Affected Products

  • lakeFS versions 1.49.1 and earlier
  • Deployments using the S3 blockstore with pre-signed multipart uploads enabled
  • Treeverse lakeFS self-hosted and managed environments prior to 1.50.0

Discovery Timeline

  • 2025-02-21 - CVE-2025-27100 published to NVD
  • 2026-06-17 - Last updated in NVD database

Technical Details for CVE-2025-27100

Vulnerability Analysis

The vulnerability resides in lakeFS's handling of S3 pre-signed multipart upload requests. An authenticated user can issue requests that cause the server to allocate excessive memory. Because allocation is not bounded relative to request parameters, repeated or crafted requests exhaust available memory and crash the process. The result is a denial-of-service condition that disrupts data access for all repository users. The flaw is classified under [CWE-400] Uncontrolled Resource Consumption.

Root Cause

The root cause is unbounded resource allocation in the S3 blockstore's pre-signed multipart upload code path. The server trusts caller-influenced parameters when sizing in-memory structures. Without upper bounds or quota enforcement, a single authenticated client can drive memory usage until the process is killed by the operating system.

Attack Vector

Exploitation requires network access to the lakeFS API and valid user credentials. The attacker submits requests that invoke the pre-signed multipart upload workflow. Repeated invocation inflates server memory until the daemon crashes. No user interaction or elevated privileges are required beyond standard authenticated access.

go
// Patch excerpt from cmd/lakectl/cmd/abuse_random_delete.go
// generate randomly selected keys as input
generator.Setup(func(add stress.GeneratorAddFn) {
    for i := 0; i < amount; i++ {
        // rand.Intn is good enough for abuse.
        //nolint:gosec
        add(keys[rand.Intn(len(keys))])
    }
})
// Source: https://github.com/treeverse/lakeFS/commit/3a625752acdf3f8e137bec20451e71d0f9fa82f2

See the GitHub Security Advisory GHSA-j7jw-28jm-whr6 for full remediation details.

Detection Methods for CVE-2025-27100

Indicators of Compromise

  • Sudden spikes in lakeFS process resident memory followed by out-of-memory termination
  • Repeated authenticated API calls invoking S3 pre-signed multipart upload endpoints from a single principal
  • Unexpected service restarts or crash loops in the lakeFS container or systemd unit

Detection Strategies

  • Baseline normal lakeFS memory consumption and alert on deviations that correlate with specific user identifiers
  • Correlate authenticated API audit logs with process health metrics to identify abusive callers
  • Review access logs for high-frequency pre-signed multipart upload requests from a single user or token

Monitoring Recommendations

  • Enable lakeFS audit logging and ship events to a centralized SIEM for behavior analysis
  • Monitor host-level OOMKilled events and container restart counts for lakeFS workloads
  • Track per-user request rates against the S3 blockstore endpoints and alert on anomalous bursts

How to Mitigate CVE-2025-27100

Immediate Actions Required

  • Upgrade lakeFS to version 1.50.0 or later across all deployments
  • If upgrade is not immediately possible, disable S3 pre-signed multipart uploads using the documented configuration flags
  • Review authenticated user accounts and revoke credentials that are no longer required

Patch Information

The fix is included in lakeFS version 1.50.0. Review the upstream patch commit 3a625752 and the GitHub Security Advisory GHSA-j7jw-28jm-whr6 for complete details.

Workarounds

  • Set the environment variable LAKEFS_BLOCKSTORE_S3_DISABLE_PRE_SIGNED_MULTIPART to true
  • Alternatively, configure the disable_pre_signed_multipart key to true in the lakeFS config YAML
  • Restrict API access using network-level controls until the patched version is deployed
bash
# Workaround via environment variable
export LAKEFS_BLOCKSTORE_S3_DISABLE_PRE_SIGNED_MULTIPART=true

# Workaround via config.yaml
# blockstore:
#   s3:
#     disable_pre_signed_multipart: true

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.