This post is part of mainframe diagnosis series. Today we will discuss about diagnostic Traces. This part 3 of the series. I am in the process of preparing for an IPCS presentation and I am refreshing some of those concepts. Thought to share this along the way.
Trace prints a trail of entries about the processing of a program or transaction. There are 6 traces that records various information regarding the processing. Some traces are overview and some are very specific. Depending on the problem you are chasing, you may have to use one or more of these traces. Below is the list.
- System Trace
This trace contains hardware and software events information. System trace usually starts during system start up and runs all the time. It maintanince traces from each processor.
Most of the dumps contains this trace data. Use IPCS SYSTRACE command to format this trace.
- Master Trace
As the name implies this is the master of traces. It contains the list of all recently issued system messages. Master Trace(MT) is started automatically at system start up but you can trun ON or OFF with TRACE command.
MT contains the messages from the hardcopy log. Traces traisl are recored in FIFO order.
- Componenet Trace
Componenet Trace (CT) records data events occurning in MVS components. Typical MVS components that you can use CT with are: BCpii, XCF, JES, RSM, WLM etc. In general CT can be used while recreating the problem.
- Transaction Trace
As the name implies, it provides traces data of a transaction. This trace can be handy as it tracks a flow of work from component to component either in single system or sysplex.
- Generalized Trace facility(GTF)
GTF is an MVS supplied service aid that tracks system and program problems. GTF tracks a wide variety of system and program events. You can control GTF options using SYS1.PROCLIB. Use START GT command to start this trace. This trace contains pretty much system trace with short data. However IBM suggest that you use GTF if you are collecting traces for a problem, say for example tracking just few events. The reason GTF provides much more detailed in depth traces for these events that are really useful in debugging the problem. You can use both GTF and SYSTEM traces.
- GFS Trace
GFS as name imples collects the data about Getmain, Freemain or Storage. This trace reports on virtually storage and can be used to locate heavy users.
reference: MVS knowledge Center