The following document outlines what is needed to
troubleshoot a JVM that causes the CPU utilization to reach 100%, or an
unusually high percentage, on a Solaris system.
If you have already contacted support, continue on to the
component-specific MustGather information. Otherwise, click: MustGather:
Read first for all WebSphere Application Server products.
100% CPU Usage specific MustGather information
Follow these instructions to set up the failing System for capturing
required information:
- Enable
verbosegc for the failing application server. This is necessary for
debugging High CPU issues and does not have significant impact on
performance.
- Clear all application server log files before starting test. You may
have to stop the Application Server to delete the files and then start
them.
- After starting the Application Server, run the following netstat and
vmstat commands to capture typical system CPU and connection activity:
% netstat -an > netstat_normal.out
% vmstat 5 12 > vmstat_normal.out |
|
Collect the following information when the application server process is
experiencing the problem:
- Execute the following script (see attached) where PID is the
process ID of the Java process causing high CPU.
- Collect the following files:
- For WebSphere V6.0:
- All files in the following directory:
install_root/profiles/profile_name/logs/server_name |
|
- A copy of server.xml located in the following
directory:
install_root/profiles/profile_name/config/cells/
cell_name/nodes/node_name/servers/server_name |
|
- For WebSphere V5.0 and V5.1:
- All files in the following directory:
install_root\logs\server_name |
|
- A copy of server.xml located at the following
directory:
install_root\config\cells\cell_name\nodes\
node_name\servers\server_name |
|
- For WebSphere V3.5, and V4.0:
- All files in the following directory:
- A copy of XMLExport for the server configuration.
- If the Web server is remote, send the
trace.log.Web_server_name.date
(V3.5), native.log (V4.0), or http_plugin.log (V5.0 and
V6.0) file from the Web server system.
- All of the *.out files created by the wascpu.sh
script.
- netstat_normal.out and vmstat_normal.out
(if possible).
- Follow instructions to Submitting
Diagnostic Information to IBM Technical Support for Problem
Determination.
If the thread dumps are not getting generated for kill -3, you may want
to try using the alternate threading library, see Changing
to the alternate threading library on Solaris 8.
For a listing of all technotes, downloads, and educational materials
specific to the 100% CPU Usage component, search the WebSphere
Application Server support site. |