Forum Discussion
Tiago_93874
Nimbostratus
NimbostratusMonitor HTTP for Oracle Application Server 10g application server
Hi, I have a scenario with two redundant application servers and two Load Balancers LTM 1500 initialy I configured a health monitor that simply test if the ports are open and this monitor work properly but they don't test if the application is available so I tried to configure the a HTTP health monitor as can be seen next:
Name: HTTP_Monitor
type: HTTP
interval:5 seconds
timeout:16 seconds
send string: GET /SIVHACTUAL/Login HTTP/1.1
Rceive string: 200
Name: HTTP_Monitor2
type: HTTP
interval:5 seconds
timeout:16 seconds
send string: GET /SIVHACTUAL/Login HTTP/1.1
Rceive string: Login
Both of this monitors appear green with this configuration but when we simulate the application failure the monitor remains green. Why the monitor with this configuration don't detect the application failure? How I configure the HTTP monitor to detect the application failure?
The one of the applications use OC4J.
This is an Oracle 10g application server.
Regards.
- Hi,
The "200" your monitor is seeing in the reply probably comes from the date as in 2008.
Try to be more specific in your match. "200 OK" for example. 
taro_112311Is the page not a dynamic page but a static page?
I think a popular method to make the page for the monitor to which the application server returns a specific string, such as "server running". (And "server running" is set as Receive String.)
Kent_Perrier_50heh, I went through this.
You need to kill the web server on the OAS machine to cause the node to be marked down. OAS is trying to do the right thing. Lets say you have your 2 servers. On server 1, you have OHS1 and OC4J1 and on server 2 your have OHS2 and OC4J2. You kill OC4J1 to simulate a problem with that server. OHS1 sees that OC4J1 is down and sends the requests to OC4J2 for processing. I am not an OAS admin, but I thing our OAS admins turned on local affinity so that OHS will talk to the local OC4J container preferably.
