Bug 2487505 (CVE-2026-11873) - CVE-2026-11873 pki-core: dogtag-pki: empty request to Dogtag /ca/rest/certrequests causes HTTP 500, java exception, and stacktrace disclosure
Summary: CVE-2026-11873 pki-core: dogtag-pki: empty request to Dogtag /ca/rest/certreq...
Keywords:
Status: NEW
Alias: CVE-2026-11873
Deadline: 2026-09-10
Product: Security Response
Classification: Other
Component: vulnerability
Version: unspecified
Hardware: All
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Product Security
QA Contact:
URL:
Whiteboard:
Depends On: 2526743
Blocks:
TreeView+ depends on / blocked
 
Reported: 2026-06-10 12:28 UTC by OSIDB Bzimport
Modified: 2026-09-01 11:33 UTC (History)
16 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed:
Embargoed:


Attachments (Terms of Use)

Description OSIDB Bzimport 2026-06-10 12:28:55 UTC
Description
An Apache-proxied Dogtag CA REST endpoint exposed by IdM (POST /ca/rest/certrequests) returns HTTP 500 with internal Java stack traces for unauthenticated malformed requests. The same unauthenticated error path emits large multi-line stack traces into the CA debug log, enabling a log-amplification resource exhaustion vector (disk growth and I/O contention) without authentication. Each unauthenticated malformed request causes ~12 KB and ~139 log lines of CA debug output, which can rapidly grow logs and induce I/O pressure or disk exhaustion. The pentesting team measured that 25 invalid-JSON requests increased /var/log/pki/pki-tomcat/ca/debug.2026-02-18.log by 305,852 bytes (~0.31 MB) and 3,477 lines, which is approximately 12,234 bytes (~12.2 KB) and 139 log lines per request. 
Impact
An unauthenticated adversary on the network segment who is able to reach the IdM server could trigger repeated Dogtag CA REST exceptions to harvest stack traces/internal class names and request-handling details, or amplify server-side logging to consume CPU/disk and degrade CA responsiveness, increasing the reliability of follow-on exploitation and DoS. 
Recommendations
Make the CA REST API fail safely by checking auth early, returning simple 4xx errors for bad requests (no stack traces), and tuning exception handling and logging so attacker-controlled input can’t spam large stack-trace logs (rate-limit/truncate).


Note You need to log in before you can comment on or make changes to this bug.