Known Issues

After upgrading to JDK 6 Update 12, Performance Monitor does not display applications (6802057)

After upgrading to JDK 6 Update 12 (1.6.0_12), Performance Monitor does not display any applications in the Applications window. This issue appears when any application on the system is running JDK 6 Update 12, even if Performance Monitor is running JDK 6 Update 11 or earlier.

Workaround: Uninstall JDK 6 Update 12, reverting back to JDK 6 Update 11 (1.6.0_11).

After a connection is restored, existing inactive views block new views (6794019)

When Performance Monitor loses the connection to a monitored item, it does not close any open view for that item. Later, if the connection is restored, trying to open a view of the item fails. Instead, Performance Monitor brings the existing inactive view of the item to the front.

Workaround: Close the existing inactive view, and then open a new view.

You can use the Close Inactive Views command on the Window menu to close all inactive views without having to locate and select each one.

After adding a JMX connection, the list of items under the connection takes a long time to display (6794029)

After adding a JMX connection and then expanding it in the tree pane, the items under the connection can sometimes take more than a minute to display.

Workaround: None. When a JMX connection is added, Performance Monitor sends several JMX messages to the connection to build the list of items the connection has available for monitoring. The elapsed response time of these messages depends on many factors, including: the load on the system hosting the domain or instance, the load on the domain or instance hosting the connection, and network bandwidth.

Only the domain administration server instance of a cluster displays a JDBC connection pool (6790176)

Even when a web application is available and enabled on a cluster, only the domain administration server instance of the cluster displays a JDBC connection pool. The other instances do not.

Workaround: In the GlassFish Admin Console, select the JDBC resource assoiciated with the web application. Then, use the Manage Targets command to make the resource available and enabled on the cluster.

Simultaneous startup of all cluster instances might cause monitoring to fail for some instances (6796388)

When a cluster is started using the asadmin start-cluster command or the Start Cluster command in the GlassFish Admin Console, GlassFish starts all the cluster instances simultaneously. In some cases, this simultaneous startup causes some of the instances to start without monitoring available. In Performance Monitor, such instances display "Unavailable" in their Memory and Uptime fields on the cluster view.

This issue has been observed on clusters of 8 instances with JMX over SSL, with the instances distributed on uniform hardware. It might also occur on larger clusters without SSL.

Workaround: Start the instances individually at staggered intervals. For example, using a script that starts the instances 10 to 20 seconds apart prevents this issue from occurring. (The exact delay time is hardware dependent.)

Using the Profiler adversely affects server performance (6802071)

Using the Profiler (available on local GlassFish domains and JVMs) adversely affects performance of the profiled application. When profiling is started, the profiled application is suspended while profiling instrumentation is enabled, and when profiling is stopped, the profiled application is again suspended while profiling instrumentation is disabled. While profiling is active, performance of the profiled application is reduced.

Workaround: None. Using the Profiler is not advised on production servers. For preproduction and development systems, here are some tips:

• If you want to profile a web application (not the entire server), use the NetBeans Profiler with its GlassFish integration instead of the Profiler in Performance Monitor. The NetBeans Profiler automatically constrains the profiling instrumentation to the application, so overall server performance is not as greatly affected.

• If you need to profile the server, try to define small rootsets (method entry points) so that only parts of the server get instrumented for profiling. Also, make as much memory and CPU time as possible available to the profiled server and to Performance Monitor.

Online help is not available when there is no internet connection (6802082)

When running Performance Monitor on a system that does not have an internet connection, online help (available at http://wikis.sun.com/display/PerfMonitorOLH) is inaccessible.

Workaround: None.

On Windows Vista, Performance Monitor fails to run after being unzipped using cygwin (6790266)

On Windows Vista, after downloading Performance Monitor and unzipping it using the cygwin command shell, Performance Monitor fails to run. Instead, Windows Vista displays this message:

Windows cannot access the specified device, path or file. You may not have the appropriate permissions to access the item.

Workaround: Do not use the cygwin command shell to unzip Performance Monitor.

On Windows XP, local GlassFish domains might fail to start when both Performance Monitor and Cisco VPN are running (6798749)

On Windows XP, if both Performance Monitor and Cisco VPN are running, attempts to start a local GlassFish domain might fail, producing a series of exceptions.

Workaround: Follow these steps:

1. Exit Performance Monitor.

2. Stop any residual Java VMs related to the failed server startup.

3. Start the GlassFish domain.

4. Wait for server startup to complete.

5. Start Performance Monitor.