- RedHat 6 - Apache 2.4.18 - Dell Boomi Status Code 503 Atom is shutting down
I have an Apache httpd24 web server running mod_proxy and mod_proxy_balancer. My test includes 2 worker nodes each running a Dell Boomi Molecule node. The goal is to have a node shutdown and flip the init status to "Err" once a 503 Error is returned. When a Boomi Atom is sent a shutdown request, if a process is running on the node it will let the process complete before the Atom is shut down. During this "wait" time any incoming web service request will be returned a 503 Error and still attempt to use the node in the shutdown state making the error responses pile up. After some research I see that the failonstatus directive can be used, however after implementation the failonstatus doesn't seem to turn off the worker. Is this a potential bug or configuration issue on my side? Any assistance to trap the 503 error would be greatly appreciated:
ProxyPass /balancer-manager ! ProxyPass / balancer://Foo_2.4/
Example of response from url during the time the Boomi Atom is stopping: Screen Shot of 503 Error Returned when trying to visit page.
Not sure if the following patch addresses this area - PR 59869
mod_proxy: Correctly consider error response codes by the backend when processing failonstatus. PR 59869 [Ruediger Pluem]