← Back to Feed CACHED · 2026-07-10 09:31:32 · CACHE_KEY tenable:87679
tenable:87679 · CWE-119 · Disclosed 2014-09-09

VMware ESXi Multiple DoS (VMSA-2014-0008)

VMware ESXi 5.0/5.1/5.5 glibc regex and getaddrinfo DoS (VMSA-2014-0008)

ASSESSED — NOISGATE
Vendor
—
—
—
Reassessed
—
—
—
Verdict: —
Do you agree?
01 · The Real Story

A twelve-year-old glibc parser bug on a hypervisor line that Broadcom stopped shipping patches for half a decade ago

VMSA-2014-0008 bundles third-party library updates for VMware vSphere. The two CVEs that produce the DoS finding — CVE-2013-0242 and CVE-2013-1914 — are glibc bugs: a buffer overflow in extend_buffers() in the regex engine, and a stack overflow in getaddrinfo() when parsing crafted hostnames/IPs. On an ESXi host these are reachable only if some management-plane service on the hypervisor actually feeds attacker-controlled input into regcomp/regexec or getaddrinfo. Affected builds are ESXi 5.0, 5.1, and 5.5 missing patch ESXi550-201409101-SG / ESXi510-201412101-SG; the CVSSv2 vector is AV:N/AC:L/Au:N/C:N/I:N/A:P (5.0, availability-only).

The vendor's MEDIUM rating was defensible in 2014 for a network-reachable, unauthenticated DoS against a hypervisor management daemon. In 2026 it is not. ESXi 5.x reached end-of-general-support in 2020 and end-of-technical-guidance in 2022; any host that still shows this plugin firing has vastly bigger problems than a glibc DoS — it is missing five years of ESXi CVEs including multiple pre-auth RCEs (Log4Shell in vCenter, OpenSLP CVE-2021-21974, CVE-2024-37085 domain-join).

"A 2013-era glibc DoS on an end-of-life ESXi 5.x hypervisor. If you still run 5.x, that's the emergency — not this CVE."
02 · The Attack Path

3 steps from start to impact.

STEP 01

Reach the ESXi management network

The attacker needs L3 reachability to the ESXi management interface (vmk0) — typically TCP/443 (hostd/rhttpproxy) or TCP/902 (authd). In any competent enterprise this is an isolated management VLAN, not internet-facing. Public Shodan/Censys sweeps for esxi mostly return residential lab hosts.
Conditions required:
  • Routable path to ESXi management VLAN
  • Firewall rules permit inbound to vmk0
Where this breaks in practice:
  • Management interfaces are almost never internet-exposed in mature shops
  • Segmentation policy typically restricts vmk0 to jump hosts
  • vSphere HCL guidance has recommended isolation since 5.0
Detection/coverage: Nessus plugin 87679 authenticated check; Tenable/Qualys ESXi credentialed scans
STEP 02

Deliver a malformed hostname or regex to a vulnerable code path

Trigger either CVE-2013-0242 (crafted multibyte regex fed to regcomp/regexec) or CVE-2013-1914 (crafted hostname/IP fed to getaddrinfo). Neither is directly reachable from the outside — the attacker must find a management-plane feature that parses attacker-supplied text as a regex or resolves an attacker-supplied hostname. No public weaponized PoC exists against ESXi.
Conditions required:
  • A network-facing hostd/vpxa feature that funnels input to glibc regex or resolver
Where this breaks in practice:
  • No public PoC mapping the glibc bugs to a specific ESXi RPC
  • Most exposed surfaces (SOAP API, DCUI, hostd) require auth
Detection/coverage: Nothing meaningful; DoS attempts appear as service crashes in vmkernel.log
STEP 03

Crash the hostd/vpxa worker

Successful trigger crashes the management daemon, not the hypervisor itself. Running VMs continue to execute — you lose vCenter management, not workloads. The daemon respawns via watchdog within seconds to minutes.
Conditions required:
  • Overflow reaches an exploitable stack/heap layout
Where this breaks in practice:
  • hostd watchdog restarts the process automatically
  • No VM guest impact, no data disclosure
Detection/coverage: vmkernel.log 'hostd stopped responding'; vCenter shows host as Not Responding briefly
03 · Compensating Control

1
Confirm ESXi management interface is isolated — Restrict vmk0 to a dedicated management VLAN reachable only from jump hosts / PAM. This nullifies the entire attack path for this CVE and every other network-reachable ESXi bug. Per noisgate LOW guidance there is no mitigation SLA — treat as backlog hygiene during the next quarterly network review.
2
Migrate off ESXi 5.x entirely — 5.0/5.1/5.5 are past end-of-technical-guidance. If plugin 87679 fires you almost certainly also have unpatched OpenSLP, log4j, and CVE-2024-37085 exposure. Plan an upgrade to 7.0 U3 or 8.0 U3 within the noisgate remediation SLA of 365 days.
3
Enable ESXi lockdown mode — Normal or Strict lockdown restricts direct hostd API access to vCenter, shrinking the population of principals who can even attempt to reach the vulnerable parsers.
4
Monitor vmkernel.log for hostd restarts — Alert on repeated hostd stopped responding events — this is your DoS detection signal since network-level IDS won't catch glibc-level triggers.
What doesn't work
  • Edge WAF / IPS — the management interface should not traverse a WAF; if it does, your architecture problem is bigger than this CVE.
  • EDR on the guest VMs — the bug lives in hostd on the hypervisor, not in guest OS user space.
  • Disabling SSH on ESXi — this DoS path is via hostd/rhttpproxy on TCP/443, not SSH.
04 · Intelligence Metadata

The supporting signals.

In-the-wild exploitationNone observed. No CISA KEV entry, no Mandiant/CrowdStrike/Unit42 campaign reporting for CVE-2013-0242 or CVE-2013-1914 against ESXi.
Public PoCglibc-level PoCs exist in academic form but no weaponized ESXi exploit is public. No Metasploit / Nuclei / ExploitDB module targets this VMSA.
EPSSBoth CVEs sit at <0.5% probability, sub-40th percentile — noise-floor forecast.
KEV statusNot listed in CISA Known Exploited Vulnerabilities catalog.
CVSSCVSS2#AV:N/AC:L/Au:N/C:N/I:N/A:P — score 5.0, availability-only, no confidentiality/integrity impact.
Affected versionsESXi 5.0, 5.1, 5.5 missing patches ESXi500-201412101-SG, ESXi510-201412101-SG, ESXi550-201409101-SG. 6.x/7.x/8.x are not affected.
Fixed versionsESXi 5.5 patch ESXi550-201409101-SG (Sep 2014); ESXi 5.1 patch ESXi510-201412101-SG (Dec 2014); ESXi 5.0 patch ESXi500-201412101-SG (Dec 2014).
Exposure telemetryShodan shows ~thousands of internet-facing ESXi hosts total; 5.x share is residual (<3%) and shrinking. Most orgs are on 6.7/7.0/8.0 by 2026.
Disclosureglibc bugs disclosed 2013-01-29; VMware advisory issued 2014-09-09, updated 2014-12-04.
Reported byglibc upstream (Bugzilla 14134, 15078); rolled into VMSA-2014-0008 by VMware Security Response.

Sources.

  1. Tenable plugin 87679
  2. Broadcom / VMware VMSA-2014-0008
  3. NVD CVE-2013-0242
  4. NVD CVE-2013-1914
  5. CISA KEV Catalog
  6. FIRST EPSS
  7. VMware Product Lifecycle Matrix
  8. Vulners VMSA-2014-0008
05 · The Call

Final Verdict
↓ DOWNGRADED to LOW (2.5/10)

Why this verdict

  • Impact ceiling is DoS-only. C:N/I:N/A:P means best case for the attacker is crashing a management daemon that auto-restarts — no guest VM compromise, no hypervisor escape, no data.
  • No exploitation signal. Not on KEV, EPSS < 1%, no public weaponization, no threat-actor reporting in twelve years.
  • Affected surface is end-of-life. ESXi 5.0/5.1/5.5 hit end-of-technical-guidance in 2020–2022; any host still firing this plugin is exposed to dozens of newer, higher-severity CVEs (OpenSLP CVE-2021-21974, CVE-2024-37085) that dwarf a glibc DoS.
  • Role multiplier — hypervisor role, chain does NOT succeed. ESXi is a canonical CRITICAL-floor deployment role, BUT the documented chain terminates at *hostd crash*, not guest escape, not vmkernel compromise, not vCenter takeover. The role multiplier does not fire because the blast radius is bounded to a temporary management-plane outage.
  • Friction stacks: requires L3 access to an isolated management VLAN, requires a reachable code path funnelling attacker input into glibc regex/resolver, and no such path has been publicly demonstrated on ESXi.

Why not higher?

MEDIUM would require either a plausible RCE path or meaningful availability impact on guest workloads. Neither exists — hostd restarts automatically and VMs keep running. There is no KEV listing, no EPSS lift, and no public PoC to justify holding at MEDIUM.

Why not lower?

Not IGNORE because a management-plane DoS on a hypervisor still has operational cost during an outage, and organizations still running ESXi 5.x should recognize this plugin as a proxy indicator for a much larger patch-hygiene failure. Document it, don't dismiss it silently.

06 · Verification

Crowdsourced verification payload.

Run this on the ESXi host itself via SSH as root, or adapt for esxcli over a vCenter proxy. Example: ssh [email protected] ./check_vmsa_2014_0008.sh. Requires shell access to the hypervisor.

noisgate-verify.sh
BASHREAD-ONLYSAFE
#!/bin/sh
# check_vmsa_2014_0008.sh - detect ESXi builds missing VMSA-2014-0008 patches
# Run on target ESXi host as root.
# Exit: 0 = PATCHED, 1 = VULNERABLE, 2 = UNKNOWN / not applicable

set -u

if ! command -v vmware >/dev/null 2>&1; then
  echo 'UNKNOWN: vmware CLI not present — is this an ESXi host?'
  exit 2
fi

VER=$(vmware -v 2>/dev/null | awk '{print $3}')
BUILD=$(vmware -l 2>/dev/null | awk '{print $NF}')
echo "Detected ESXi version=${VER} build=${BUILD}"

case "$VER" in
  5.0*)  FIX_BUILD=2312428 ;;   # ESXi500-201412001
  5.1*)  FIX_BUILD=2323231 ;;   # ESXi510-201412001
  5.5*)  FIX_BUILD=2068190 ;;   # ESXi550-201409001
  6.*|7.*|8.*) echo 'PATCHED: ESXi 6.x/7.x/8.x not affected by VMSA-2014-0008'; exit 0 ;;
  *) echo 'UNKNOWN: unrecognized ESXi major version'; exit 2 ;;
esac

if [ -z "$BUILD" ] || ! echo "$BUILD" | grep -qE '^[0-9]+$'; then
  echo 'UNKNOWN: could not parse build number'
  exit 2
fi

if [ "$BUILD" -ge "$FIX_BUILD" ]; then
  echo "PATCHED: build ${BUILD} >= fix build ${FIX_BUILD}"
  exit 0
else
  echo "VULNERABLE: build ${BUILD} < fix build ${FIX_BUILD} — apply VMSA-2014-0008 patch"
  echo "NOTE: ESXi 5.x is end-of-support. Priority action is upgrade to 7.0 U3 or 8.0."
  exit 1
fi