Static vs. dynamic tracing, Tracing and debugging – HP Integrity NonStop H-Series User Manual
Page 206
Problem Resolution
HP NonStop AutoTMF Software User’s Guide—429952-013
C-8
Tracing
Remember that tracing is a diagnostic tool. As such, tracing can generate large
amounts of data. Tracing should be used with caution in any production environment
and not at all in performance critical environments.
Tracing is initiated in two ways:
•
command.
•
DEFINEs
You can trace a single process or all instances of a running program. Up to eight such
traces, numbered 0 to 31, can be active simultaneously.
Static vs. Dynamic Tracing
You can trace a single process or a set of programs. Tracing can be either static or
dynamic:
•
Static tracing starts at the time the process is launched. Tracing is initiated by the
monitor process through the CI TRACE command.
Tracing can also be started by specifying one of three DEFINES in the program
startup command stream or the Pathway server configuration.
•
Dynamic tracing can be initiated at any time before or during program execution by
the monitor through the CI TRACE command.
You can enable dynamic tracing on your system with the
command to set the global DYNAMICTRACE to ON.
The global setting DYNAMICTRACE can be overridden for a single process by
using DEFINEs. See
and
for details.
Note that the dynamic tracing global or DEFINE must be specified when a process
is started, which tells the process to check for tracing at various times.
Allowing dynamic tracing increases overhead, so use dynamic tracing judiciously.
Tracing and Debugging
Tracing is often done in conjunction with the debugging of a program using INSPECT.
In addition to the standard TACL and Pathway DEBUG options, you can specify that a
program should be started in INSPECT with the CI TRACE command. Tracing can be
particularly useful when debugging Pathway servers since tracing does not require
modifying the server configuration.