---
id: "CVE-2026-115817"
kind: "vulnerability"
aliases: ["CVE-2026-115817"]
title: "Starman: stalled request body holds every Perl worker, causing denial of service"
severity: "medium"
cvss: null
epss: null
exploited: false
status: "open"
category: "package"
products: []
ecosystems: []
tags: ["package", "perl", "psgi", "self-hosted", "starman"]
not_affected_if: ["proxy-with-body-timeouts", "version-0.4019-or-later"]
published_at: "2026-10-25T17:17:00+00:00"
first_seen_at: "2026-10-25T17:42:01+00:00"
updated_at: "2026-10-25T17:43:53+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-115817", "first_seen_at": "2026-10-25T17:42:01+00:00"}]
---

# Starman: stalled request body holds every Perl worker, causing denial of service

Starman, a Perl PSGI web server, reads request bodies with blocking reads and no deadline, so read_timeout only covers headers. A remote client that sends complete headers and then stalls the body can hold a worker indefinitely, and a few such connections exhaust the default worker pool.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- Check the installed Starman version with `perl -MStarman -e 'print $Starman::VERSION'` or `cpanm --info Starman`
- Confirm whether the server is exposed to untrusted clients directly rather than behind a proxy with body timeouts (unknown from the sources)

## Fix

- Upgrade Starman to `0.4019` or later, which applies read_timeout to each body read and adds body_timeout (default 300 seconds)
- Until upgraded, put a reverse proxy with request body timeouts in front of Starman (unknown from the sources)

## Timeline

- 2026-10-25T17:42:01+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-115817

## From the sources

Starman versions before 0.4019 for Perl allow a stalled request body to hold a worker indefinitely, as read_timeout covers only the headers.

Content-Length and chunked request bodies are read with blocking reads and no deadline. A client that sends complete headers and then stalls or trickles the body keeps the worker waiting until it disconnects.

A remote attacker can take every worker with a few such connections, five with the default worker count, so the server stops answering requests.

Version 0.4019 applies read_timeout to each wait for body data and adds a body_timeout limit, 300 seconds by default, on the total time to read a body.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_