You can manually start a memory leak analysis of a server by using the WebSphere Automation UI. Processes on the monitored server pause while the
memory heap dump is generated for the health investigation, which might cause applications that are
running on the server to be affected, possibly impacting application users.
Before you begin
You must have a user profile with the Manage investigations
permission. For more information, see Configuring roles and
permissions.
Procedure
Log in to WebSphere Automation; in the menu, click Operate > Application runtimes.
Open the Server management page or the Server
details page for the server that you want to analyze.
Figure 1. Viewing Server management page with an inventory of managed application
servers
If you do not see the menu option Operate > Application Runtimes, or if you see a message that you are not authorized, then you do not have permission
to access the page. For more information about permissions, see Roles and
permissions. If you do not see a listing of servers, then you need to register your servers
with WebSphere Automation. For instructions on registering servers with
WebSphere Automation, see Registering a server. Figure 2. Viewing Server details page with information about health investigations
for a managed application server
The Start memory analysis on
server dialog opens.
If you are on the Server management page, locate the server whose
memory you want to analyze in the list, click its action menu (the three-dot icon) and select
Start memory analysis. If you are on the Server details page, select the
Health investigations tab and then click Start memory
analysis.
In the Start memory analysis on server dialog, click
Start memory analysis.
A manually initiated memory analysis
involves the same memory heap dump collection procedure and memory heap dump analysis as an
automatically initiated health investigation.
If you started the memory analysis from the Server management page action
menu, the Health page opens. A notification indicates that a new health
investigation is created. The new health investigation is in the first row of the table and has a
status of Investigating.
If you started the memory analysis from the Health investigations tab on
the Server details page, the new investigation number appears under
Memory leak investigations.
If there is a problem, work with application owners to resolve the issue. To fix
memory leaks, application owners need the information in the archive that you download from the
investigation details view after the investigation is complete. The files in the archive include:
A summary of the investigation.
A heap dump that was collected when the leak was detected.
Files that were generated during the analysis of the heap dump.
Log files from the investigation steps.
The summary file might provide enough information to identify the location of the memory leak
in their application. However, if further analysis is needed, the application owners can analyze the
heap dump by using tools such as the Eclipse Memory Analyzer Tool with DTFJ and IBM
Extensions. The Eclipse Memory Analyzer Tool with DTFJ and IBM Extensions can use the
files that are generated during the initial analysis of the heap dump to greatly reduce the amount
of time that is needed to open the heap dump file. Those analysis files are located in the same
directory as the heap dump file in the archive with the investigation results.
