Problem Insights server task is not responsive, or Problem Insights server process terminates without a clear cause
If the Problem Insights server task is not responsive, or if the Problem Insights server process terminates without a clear problem cause in its logs, the reason might be that the Problem Insights server generated so many messages that the Job Entry Subsystem (JES) Spool SYSOUT limit is exceeded.
This problem occurs only if the Problem Insights server is run as a started task on IBM® z/OS® UNIX System Services. An indication of this problem is that you might see the following type of output in the z/OS system log:
$HASP375 GLAPISRV ESTIMATED KBYTES EXCEEDED
$HASP395 GLAPISRV ENDED - ABEND=S722
Cause
This problem is more likely to occur if the following properties have a value of
info or finest:
- loglevel property (logging level) in the cli.config file
- traceSpecification property (tracing level) in the server.xml file
However, the main cause is based on whether the JES Spool SYSOUT limit is set, and what that limit value is. For more information about the JES Spool SYSOUT limit, see IEFUSO — SYSOUT Limit Exit.
The longer that the Problem Insights server runs, the more likely it is to exceed a specified SYSOUT limit.
Solution
Messages that are written to the STDOUT and STDERR DD names in the Problem Insights server task
are also written to the messages.log and trace.log files
in the Problem Insights server directory on the IBM z/OS UNIX System Services file system, for
example:
PI_HOME/wlp/usr/servers/piFrameworkServer/logs
Therefore, you can manually suppress the writing of messages to the STDOUT and STDERR DD names
without a risk of data loss. To prevent Problem Insights server availability problems, manually
suppress the writing of these messages if your environment enforces a restrictive SYSOUT limit. To
suppress the writing of these messages, update the Problem Insights server task to redirect the
output to DUMMY.
- Example
-
//STDOUT DD DUMMY //STDERR DD DUMMY