failonstatus does not appear to induce failover sometimes
Issue
-
We have
failonstatus
set on our mod_proxy_balancer to force error states and cause requests to failover to other balancer members in the event of certain response codes from the backend. Sometimes we don't see failover as expected from this setting. We also note messages like the following in our error log:[error] proxy: BALANCER: (balancer://mycluster). Forcing recovery for worker (http://failedworker.com/)
Environment
- Red Hat Enterprise Linux
- JBoss Enterprise Web Server (EWS)
- Apache httpd
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.