(ZOS) Formatting CTRACE data in batch mode with IPCS

Use the interactive problem control system (IPCS) in batch mode to automate formatting CTRACE data.

We must create an IPCS dump directory before we can use IPCS in batch mode. When setting up IPCS, the installation may customize IPCS for its users. This customization can include modifying the IBM-supplied BLSCDDIR CLIST with default values for creating an IPCS dump directory.

To view messages or application trace data from Component Trace, use the interactive problem control system (IPCS) to format the data. Using IPCS in batch mode is the easiest method of formatting data, especially if we do not have much experience with using IPCS, TSO/E and ISPF. Through batch mode, we can use IPCS to format trace data and write it to an MVS™ data set. Optionally, you may copy the contents of that data set into an HFS file for viewing.

When the installation has modified the BLSCDDIR CLIST the steps outlined herein will create an IPCS dump directory.

1. Decide on a fully-qualified data set name for the directory.

2. From the TSO/E command prompt, enter the BLSCDDIR command, specifying the data set name.

   For example, to create a dump directory named IBMUSER.DDIR, enter:

   %blscddir dsn('ibmuser.ddir')
   

If the installation has not customized IPCS, we might need to alter other BLSCDDIR CLIST parameters. See the z/OS MVS IPCS User's Guide, SA22-7596 and z/OS MVS IPCS Commands, SA22-7594 for more details about using the BLSCDDIR CLIST to create a dump directory.

Perform the following steps to use IPCS in batch mode to format application trace data:

Tasks

1. Create a file and copy the following sample JCL into it. This JCL invokes IPCS to extract and format JRAS trace data and write it into an MVS data set, and then uses the TSO/E OPUT command to copy the formatted data from the MVS data set into an HFS file.

   //IBMUSERX   JOB,
   // CLASS=J,NOTIFY=&SYSUID,MSGCLASS=H
   //IPCS       EXEC PGM=IKJEFT01,REGION=4096K,DYNAMNBR=50
   //IPCSDDIR   DD DSN=IBMUSER.DDIR,DISP=SHR
   //IPCSDOC    DD STDERR=H
   //JRASTRC    DD DSN=IBMUSER.CB390.CTRACE,DISP=SHR
   //IPCSPRNT   DD DSN=IBMUSER.IPCS.OUT,DISP=OLD
   //SYSTSPRT   DD STDERR=*
   //SYSTSIN    DD *
   IPCS
   DROPDUMP DDNAME(JRASTRC)
   PROFILE LINESIZE(80)PAGESIZE(99999999)
   SETDEF NOCONFIRM
   CTRACE COMP(SYSBBOSS) DDNAME(JRASTRC) FULL PRINT +
          NOTERMINAL
   DROPDUMP DDNAME(JRASTRC)
   END
   /*
   //OPUT       EXEC PGM=IKJEFT01,REGION=4096K,DYNAMNBR=50
   //SYSTSPRT   DD STDERR=*
   //SYSTSIN    DD *
   oput 'ibmuser.ipcs.out' '/u/ibmuser/ipcs/jrastrace.txt' TEXT
   /*
   

2. Edit the sample JCL to replace IBMUSER.DDIR with the data set name that we used for the IPCS dump directory we created.

   1. Use the PAGESIZE parameter on the PROFILE statement only if we do not want to print the output data set.

   2. You may replace the HFS file name with the name of an existing HFS file, but we do not have to do so. The OPUT command processing will create a new HFS file, if the one specified does not exist, and grants read and write access to that file for your user ID only.

      If we do specify an existing HFS file, the OPUT command processing will write over any data already in that file. To know more about the OPUT command, see the z/OS UNIX System Services Command Reference, SA22-7802.

   3. Change the data set name specified on the JRASTRC DD in the example to the name of the data set containing the CTRACE data.

   4. Change the name of the MVS data set on both the JRASTRC DD statement and the OPUT command in the SYSTSIN stream, as necessary. The formatted output of the JRAS CTRACE data is first written to the MVS data set specified by the IPCSPRNT DD statement and then (optionally) copied to the HFS data set. We must either pre-allocate this data set, or change the sample JCL to allocate the data set. This data set should have a record format of VBA and a record length of 133.

3. Submit the JCL to start the IPCS batch job.

What to do next

Once we are done we can use a UNIX editor, such as vi, to view your trace data in the HFS file. To know more about the UNIX editors, see z/OS UNIX System Services User's Guide, SA22-7801.

CTRACE enables us to view multiple traces together with the trace data from the various sources intermixed based on the time stamp. See z/OS MVS IPCS Commands, SA22-7594, for specifics on using this MERGE subcommand.

Subtopics

• Sample JCL to display WebSphere for z/OS trace data
  Use this sample JCL to display trace data for WAS for z/OS.

Formatting CTRACE data with an IPCS dialog