[HTTP] Multiple managed Web servers cause startup issue using admininstrative console

Multiple managed Web servers cause startup issue using admininstrative console

Problem(Abstract)

When multiple Web servers are managed through a single WebSphere® Application Server node, it may not be possible to start or stop all Web servers from that Application Server's administrative console.

Example: User creates two IBM® HTTP Server configurations ( webserver1 and webserver2 ). From the administrative console, it is possible to start webserver1, but it is not possible to start webserver2.

Cause

A second IBM HTTP Server Web server can not be started from the WebSphere Application Server administrative console when both Web servers have the same installation root and the first Web server uses a default PID file of
install_root/logs/httpd.pid.
The problem seems to be that the launch code in WebSphere Application Server checks status of the Web server before allowing the START command. The real contributor to this problem is that the hard coded PIDFILE logs/httpd.pid is used to check for the status of the Web server. If first Web server is running, the second Web server uses the default httpd.pid and sees the first IBM HTTP Server Web server as running, and will not start. For problem with stopping Web server, see PK07361.
install_root/conf/httpd.conf
# PidFile: The file in which the server should record its process
# identification number when it starts.

Resolving the problem

To correct this issue:

1. Stop the IBM HTTP Server, or other managed Web servers.
   
   
2. If two Web servers share the same installation root, neither Web server should use the default PID file of logs/httpd.pid. If one uses a default, change it another value such as logs/httpd1.pid.
   
   
3. Start webserver1 – this should start successfully
   
   
4. Start webserver2 – now this also starts fine.