An OHS <-> OC4J Failover Test Fails With HTTP 404 When Accessing The Remote OC4J Instance On HP-UX Itanium IA64 (Doc ID 1339569.1)

Last updated on FEBRUARY 16, 2017

Applies to:

Oracle HTTP Server - Version 10.1.2.3.0 and later
HP-UX Itanium

Symptoms

On an Application Server 10.1.2.3.0 environment on HP-UX IA64 11.31 where MOD_OC4J has been configured in HA as follows:

Oc4jSelectMethod roundrobin:local
Oc4jMount /myapp myOc4j
Oc4jMount /myapp/* myOc4j



For example: myOc4J is an OC4J instance running in two clustered midtier nodes, so Oc4jMount will failover requests to any of the alive members of the cluster. See the Oc4jMount syntax for details.

When one of the OC4J instances go down the subsequent requests fails with a HTTP 404 error, instead of getting directed to the alive node.

The issue can be reproduced at will with the following steps:

1. Have OHS and the two OC4J instances running
2. Perform several requests and verify they work fine
3. Stop the local OC4J instance and keep sending requests. An HTTP 404 error will appear in the $ORACLE_HOME/Apache/Apache/logs/access_log file:


10.207.228.11 - - [22/Jun/2011:09:45:19 +0200] "GET /myapp/myservlet HTTP/1.1" 200 1587
... <== The myOc4j instance is stopped here
10.207.228.11 - - [22/Jun/2011:09:45:19 +0200] "GET /myapp/myservlet HTTP/1.1" 404 139
10.207.228.11 - - [22/Jun/2011:09:45:21 +0200] "GET /myapp/myservlet HTTP/1.1" 404 139


The $ORACLE_HOME/Apache/Apache/logs/error_log does not show any MOD_OC4J error .

The output of http://<host>:<port>/oc4j-service?cmd=p , implemented as per <Note 334502.1> shows MOD_OC4J is detecting the remote OC4J instance as alive as expected, but all requests sent to it fail with HTTP 404 (Note is not an HTTP 500 or HTTP 503 error).

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms