If you plan to collect log data for WebSphere® Application Server for z/OS®,
you must determine the application servers from which to collect log
data.
Then, for each of these application servers, you must determine
where to retrieve the log data.
If the application server is configured to use High Performance
Extensible Logging (HPEL) mode, a best practice is to retrieve the
log data by using the HPEL API.
If the application server is configured to use basic logging, the
log data is retrieved from JES job logs, z/OS UNIX log
files, or both, depending on how the server is configured.
If the application server is configured to use HPEL
mode
For each application server that is configured to use
HPEL mode, complete the following steps:
- Use the WebSphere Integrated
Solutions Console to determine the HPEL logging and trace directories.
Logging and trace information is typically in the same directory,
but it can be configured to be in different directories.
- Determine whether logging data only, trace data only, or both,
is collected.
Logging data includes data from the java.util.logging package
(the level DETAIL and higher), the System.out stream,
and the System.err stream.
Trace data includes
data from the java.util.logging package (the
level DETAIL and lower).
- Ensure that the user ID that is associated with the z/OS Log Forwarder procedure
is authorized to read the HPEL logging and trace directories and files.
If the application server is configured to log to
JES job logs
For each application server that is configured
to log to JES job logs, complete the following steps:
- Determine which regions of the application
server to collect log data from.
| Region |
Focus area |
|---|
| Controller region |
Inbound and outbound communication, security, and transaction
control |
| Servant region |
Most of the application server components |
| Adjunct region |
Internal messaging |
- For each application server region, determine the job name.
| Region |
Typical job name |
|---|
| Controller region |
Server short name |
| Servant region |
Job name for the controller region with an "S" appended |
| Adjunct region |
Job name for the controller region with an "A" appended |
- For each application server region,
determine whether to collect SYSOUT data, SYSPRINT data, or both types
of data.
SYSOUT data includes Java™ logs
(non-trace levels) and native message logs.
SYSPRINT data includes Java logs (with trace levels) and
native trace.
If the application server is configured to log to z/OS UNIX log
files
For each application server that is configured to
log to z/OS UNIX log files, complete the following steps:
- Determine which regions of the application
server to collect log data from.
| Region |
Focus area |
|---|
| Controller region |
Inbound and outbound communication, security, and transaction
control |
| Servant region |
Most of the application server components |
| Adjunct region |
Internal messaging |
- For each application server region,
determine whether to collect SYSOUT data, SYSPRINT data, or both types
of data.
SYSOUT data includes Java logs
(non-trace levels) and native message logs.
SYSPRINT data includes Java logs (with trace levels) and
native trace.
- Determine the file path of each z/OS UNIX log file to be collected.
- For each z/OS UNIX file to be collected, determine whether
the path name is constant or varies. The path name varies in the following
situations:
- When date and time substitution is used in the file name that
is specified in the data definition (DD) statement. The use of date
and time substitution causes a new log file to be created for each
server instance.
- When the WebSphere environment
variable redirect_server_output_dir is used to
redirect output to files. The use of this variable causes a new log
file to be created for each server instance. It also gives you the
capability to use the ROLL_LOGS parameter of
the modify command to create a new set of log files.
For more information about log paths for rolling logs, see Collection of log data from a rolling log.
- Ensure that the user ID that is associated with the z/OS Log Forwarder procedure
is authorized to read the z/OS UNIX log files.