Known Problems and Workarounds

What is the problem? If you configure multiple instances of the same Netscape server type, and the ServerID of one server is a substring of the ServerID of another server, ServiceGuard may fail the wrong package unexpectedly.

What is the workaround? Choose distinct, non-overlapping instance names as in the following example:

         ServerID1:          httpd-default1
         ServerID2:          httpd-default2

Do not use httpd-default for ServerID1 and httpd-default1 for ServerID2.

What is the problem? In Netscape Enterprise Server (version 3.6), the installation directory and port setting have been modified.

What is the workaround? In the runhalt script enterprise.sh (version B.01.02) line 51, modify the environment variable SERVER as follows, to reflect the new installation directory:

        SERVER=/opt/ns-enterprise36/https-${SERVER_ID}
         

When accessing the administration server for Enterprise Server version 3.6, use port 17493. For example in your browser, use the following URL:

        http://192.1.1.3:17493

where 192.1.1.3 is the IP address of the system on which you are running the administration server.

What is the problem? In Enterprise Server version 3.6, a Unix daemon, uxwdog, now monitors the health of the web server daemon, known as https-<InstanceName>, where InstanceName is the name you give to a specific server instance. (The default daemon name is https-default.)

If the web server process terminates, the uxwdog daemon will automatically restart the web server daemon. The problem is that when the uxwdog daemon terminates with the web server still running, the administration server cannot be used to stop or start the web server. (This problem can be caused by using the command kill -9 <PID>, where <PID> is the process id of the uxwdog daemon.) Furthermore, the toolkit that is monitoring the uxwdog daemon will cause the package to fail over to the adoptive node, and it will fail to run on the adoptive node.

What is the workaround? If the uxwdog daemon fails, the web server process should be manually killed as in the following:

         kill -9 <PID> 

where <PID> is the process id of the web server process. This will allow the web server to be started through the control of the administration server.