Introduction to Debugging
The VM/370 Control Program manages the resources of a single computer
such that multiple computing systems appear to exist. Each "virtual computing system,!! or virtual machine, is the functional equivalent of
an IBM System/370. Therefore, the person trying to determine the cause
of a VM/370 software problem must consider three separate areas:
1. The Control Program (CP), which controls the resources of the real
machine.
2. The virtual machine operating system running under the control of CP, such as CMS, RSCS, OS, or 3. The problem program, which executes under the control of a virtual
machine operating Information that explains how to debug CP or CMS is contained in this
book; information explaining how to debug applications programs is in
the VMLJ70 For information that explains how to use
the VM/370 Interactive Problem Control System (IPCS) for debugging,
refer to the (IPCS) If an IPCS problem is caused by a virtual machine operating system
(other than CMS and RSCS), refer to the publications pertaining to that
operating system for specific information. However, use the CP debugging
facilities, such as the CP commands, to perform the recommended
debugging procedures discussed in the other publication.
If it becomes necessary to apply a PTF (program Temporary Fix) to a
component of VM/370, refer to the VMLJ70 for detailed information on applying PTFs,. How to Start Debugging
Before you can correct any problem
6 you must recognize that one existse Next, you must identify the problem, collect informaticn, and determine
the cause so that the problem can be fixed. When running VM/370
1 you
must also decide whether the problem is in CP, the virtual machine, or
the problem program.
A good approach to debugging is: 1. Recognize that a problem exists.
2. Identify the problem type and the area affected.
3. Analyze the data you have available, collect more data if you need
it, then isolate the data that pertains to your problem.
4. Finally, determine the cause of the problem and correct it. Part 1. Debugging with VM/370 3

DOES A PROBLEM EXIST? There are four types of problems:
1. Loop
2. Wait state
3. Abend (abnormal end)
4. Incorrect results
The most obvious indication of a problem is the abnormal termination
of a program. Whenever a program abnormally terminates, a message is
issued. Figure 1 lists the possible abend messages and identifies the
type of abend for these messages.
Message
(Alarm rings) DMKDMP9081 SYSTEM FAILURE CODE xxxxxx DMKDMP905W SYSTEM DUMP FAILURE; PROGRAM CHECK DMKDMP906W SYSTEM FAILURE; MACHINE CHECK, RUN SEREP DMKDMP907W SYSTEM DUMP FAILURE; FATAL I/O ERROR DMKCKP900W SYSTEM RECOVERY FAILURE; PROGRAM CHECK DMKCKP901W SYSTEM RECOVERY FAILURE; MACHINE CHECK, RUN SEREP DMKCKP902W SYSTEM RECOVERY FAILURE; FATAL I/O ERROR - NUCL CYL - WARM CYL DMKCKP922W SYSTEM RECOVERY FAILURE; INVALID SPOOLING DATA DMKCKP910W SYSTEM RECOVERY FAILURE; INVALID WARM START CYLINDER DMKCKP911W SYSTEM RECOVERY FAILURE; WARM START AREA FULL DMKCKS903W SYSTEM RECOVERY FAILURE; VOLID xxxxxx ALLOCATION ERROR CYLINDER xxx DMKCKS912W SYSTEM RECOVERY FAILURE; VOLID xxxxxx NOT MOUNTED DMKCKS915E PERMANENT I/O ERROR ON CHECKPOINT CYLINDER DMKCKS916E ERROR ALLOCATING SPOOL FILE BUFFERS DMKCKS917E CHECKPOINT CYLINDER INVALID; CLEAR STORAGE AND COLD START Figure 1. Abend Messages (Part 1 of 3)
Type of Abend CP abend, system dumps to
disk. Restart is automatic.
If the dump program encoun­
ters a program check, ma­
chine check, or fatal I/O error, a message is issued
indicating the error. CP enters the wait state with
code 003 in the PSW. If the checkpcint program
encounters a program check,
a machine check, a fatal I/O error, or an error relating
to a certain warm start
cylinder or warm start data
conditions, a message is
issued indicating the error
and CP enters the wait state
with code 007 in the PSW. If the checkpoint start
program encounters a severe
error, a message is issued
indicating the error and CP enters the wait state with
code OOE in the PSi.