Bug 1643560

Summary:	selinux breaks haproxy basic auth when using nbthread
Product:	[Fedora] Fedora	Reporter:	Robin <robin.bjorklin>
Component:	selinux-policy	Assignee:	Lukas Vrabec <lvrabec>
Status:	CLOSED DUPLICATE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	29	CC:	bperkins, dwalsh, lvrabec, mgrepl, mmalik, plautrba, robin.bjorklin
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-10-29 13:47:26 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Robin 2018-10-26 14:37:34 UTC

Description of problem: When running haproxy with the `nbthread` directive in the configuration selinux breaks the `http-request auth` functionality.


Version-Release number of selected component (if applicable): haproxy-1.8.14


How reproducible: Always


Steps to Reproduce:
1. git clone https://github.com/rbjorklin/selinux-haproxy-bug.git
2. ./setup29.sh
3. visit http://localhost:8080/
4. enter credentials test/test

Actual results: Credential popup keeps reappearing.


Expected results: Proceeds to loading page correctly.


Additional info: The above works when `setenforce 0` is executed. No selinux denials are created even with `semodule -DB`.
The problem also appears under Fedora 28.

Comment 1 Milos Malik 2018-10-29 12:44:56 UTC

Could you collect SELinux denials, which appeared on your machine as result of the reproducer, and attach them here?

# ausearch -m avc -m user_avc -m selinux_err -m user_selinux_err -i -ts today

The denials will help us understand where the problem is.

Thank you.

Comment 2 Robin 2018-10-29 13:47:26 UTC

I'm closing this one and setting haproxy as the target component instead of selinux-policy. Some further testing on my end showed that the problem did appear with selinux turned off after all.

*** This bug has been marked as a duplicate of bug 1643941 ***