Bug 1309598

Summary:	ProxyErrorOverride On disables workers when a 50x error code is returned by the backend server
Product:	[JBoss] JBoss Enterprise Web Server 2	Reporter:	Ingo Weiss <iweiss>
Component:	mod_cluster, httpd	Assignee:	George Zaronikas <gzaronik>
Status:	CLOSED EOL	QA Contact:	Michal Karm Babacek <mbabacek>
Severity:	high	Docs Contact:
Priority:	medium
Version:	2.1.0	CC:	aogburn, bperkins, gzaronik, jclere, jdoyle, jpallich, pslavice, rsvoboda
Target Milestone:	CR01	Flags:	dmichael: needinfo-
Target Release:	2.1.1
Hardware:	Unspecified
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:	When a VirtualHost uses ProxyPass to proxy traffic to the backend uses ProxyErrorOverride to host custom error pages on Apache httpd side, when the backend replies with a 50x error code mod_proxy/mod_cluster marks that worker as down, breaking session stickiness. This issue is fixed.	Story Points:	---
Clone Of:
Clones:	1338664 (view as bug list)		Environment:
Last Closed:	2019-06-13 09:17:32 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:
Bug Depends On:
Bug Blocks:	1338664

Description Ingo Weiss 2016-02-18 09:07:40 UTC

Description of problem:

When a VirtualHost uses ProxyPass to proxy traffic to the backend uses ProxyErrorOverride to host custom error pages on Apache httpd side, when the backend replies with a 50x error code mod_proxy/mod_cluster marks that worker as down, breaking session stickiness. This behaviour is quite similar to have failonstatus=500 for example

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

Apache httpd 2.2.26 from JBoss EAP and mod_cluster 1.2.11.Final

How reproducible:

Always

Steps to Reproduce:
1. Configure a VirtualHost with ProxyPass / balancer://balancername/
2. Add ProxyErrorOverride On
3. Add at least ErrorDocument 500 /error500.html
4. Access the backend normally and the trigger a 500 error

Actual results:

Worker is marked as down and connection is forwarded to another worker, breaking session stickiness. Client receives custom 500 error page as expected.

Expected results:

Second request not being forwarded to a different worker

Comment 3 George Zaronikas 2016-07-17 18:58:32 UTC

Backported upstream patch https://svn.apache.org/viewvc?view=revision&revision=1683112 in order to resolve the issue

Comment 5 Michal Karm Babacek 2016-08-10 23:50:23 UTC

Apparently, single HTTP 500 error code coming from a one worker doesn't break session stickiness:

>hudson@rhel7GAx86-64:~$ curl 'http://192.168.122.172:8746/clusterbench/httpcode/?http_code=200' -c cookie.txt -b cookie.txt
>Done. 
>HTTP Code was: 200
>JVM route: tomcat-7-2
>Session ID: 65F8C139571EF218315BAD6BD94E8D18.tomcat-7-2
>Session isNew: false
>
>hudson@rhel7GAx86-64:~$ curl 'http://192.168.122.172:8746/clusterbench/httpcode/?http_code=200' -c cookie.txt -b cookie.txt
>Done. 
>HTTP Code was: 200
>JVM route: tomcat-7-2
>Session ID: 65F8C139571EF218315BAD6BD94E8D18.tomcat-7-2
>Session isNew: false
>
>hudson@rhel7GAx86-64:~$ curl 'http://192.168.122.172:8746/clusterbench/httpcode/?http_code=500' -c cookie.txt -b cookie.txt
>My Custom 500 page.
>
>hudson@rhel7GAx86-64:~$ curl 'http://192.168.122.172:8746/clusterbench/httpcode/?http_code=200' -c cookie.txt -b cookie.txt
>Done. 
>HTTP Code was: 200
>JVM route: tomcat-7-2
>Session ID: 65F8C139571EF218315BAD6BD94E8D18.tomcat-7-2
>Session isNew: false
>
>hudson@rhel7GAx86-64:~$ curl 'http://192.168.122.172:8746/clusterbench/httpcode/?http_code=200' -c cookie.txt -b cookie.txt
>Done. 
>HTTP Code was: 200
>JVM route: tomcat-7-2
>Session ID: 65F8C139571EF218315BAD6BD94E8D18.tomcat-7-2
>Session isNew: false