|
|
It is not always well understood that updating an application can lead to situations where it is no longer available. Performing an update without paying special attention to things such as session replication, or even worse, restarting all application instances at the same time, might lead to downtime of the application or lost requests. The most obvious cause for downtime is a restart of an application or appserver. Example 5-1 shows that the application was not available while restarting it.
Example 5-1 Restarting an application
[4/15/05 10:10:17:864 CDT] 0000012a ApplicationMg A WSVR0217I: Stopping application: BeenThere [4/15/05 10:10:18:017 CDT] 0000012a EJBContainerI I WSVR0041I: Stopping EJB jar: BeenThere.jar [4/15/05 10:10:18:028 CDT] 0000012a ApplicationMg A WSVR0220I: Application stopped: BeenThere [4/15/05 10:10:24:646 CDT] 000000e4 ApplicationMg A WSVR0200I: Starting application: BeenThere [4/15/05 10:10:24:702 CDT] 000000e4 EJBContainerI I WSVR0207I: Preparing to start EJB jar: BeenThere.jar [4/15/05 10:10:24:717 CDT] 000000e4 EJBContainerI I WSVR0037I: Starting EJB jar: BeenThere.jar [4/15/05 10:10:24:733 CDT] 000000e4 WebGroup A SRVE0169I: Loading Web Module: BeenThere WAR. [4/15/05 10:10:24:776 CDT] 000000e4 VirtualHost I SRVE0250I: Web Module BeenThere WAR has been bound to default_host[*:9080,*:80,*:9443,wasmember06.ibmredbook.com:9080,wasmember06.ibmredbook.com:80,wasmember06.ibmredbook.com:9443,wasmember05.ibmredbook.com:9080,wasmember05.ibmredbook.com:80,wasmember05.ibmredbook.com:9443]. [4/15/05 10:10:24:852 CDT] 000000e4 ApplicationMg A WSVR0221I: Application started: BeenThere
Another cause for unavailability are changes made to the application that interfere with HA services under the control of the HAManager. If a new application version has changed session objects signatures, you are exposed to class cast exceptions (ClassCastException).
