Owner's manual

Problem Resolution
HP NonStop AutoTMF Software User’s Guide—429952-016
C-7
Tracing
To list the values of system wide parameters, use the CI command INFO GLOBALS.
To change any global parameter, use ALTER GLOBAL.
To list the AutoTMF configuration for selected audited files (SEPARATEX,
COMMONTX, and so on), use INFO ATMFFILESET. To configure files, use ADD
ATMFFILESET.
To list the AutoTMF configuration for selected programs, use INFO ATMFPROGRAMS.
To configure programs, use ADD ATMFPROGRAMS.
To verify if a specific DEFINE has been added for a program, look at a trace of the
program’s execution. All active DEFINEs are listed in the output of the trace.
Checking the State of the Monitor and MapDB
The monitor process must be up at all times. To check on the status of the monitor, use
the STATUS MONITOR command to display the monitor’s vital statistics.
To look at the configuration of the MapDB, use the INFO MAPDB command.
Checking the State of the Application Programs
A program that has not been prepared will run without warnings until the program
accesses an audited file without having an active transaction. The operation will then
fail with an error 75. To verify that a program is properly prepared, use the INFO
PROGRAM command. To display a list of unprepared programs, issue the command
on a file-set using the UNPREPARED option.
To find out who is using the runtime library, use the INFO LIBRARY command. This
command displays all the processes that have the library opened, including the CI from
which the command is issued.
Checking for Locks
Locking of audited files is different than locking of non-audited files. This could
introduce concurrency problems. To display the locks held for a file, a file set, or the
entire system, use the CI LISTLOCKS (LL)
(or LL) command. Use the DEADLOCK
option of the command to display deadlocks.
Tracing
AutoTMF offers a comprehensive set of tracing options. Virtually all file system and
TMF operations performed by a program can be traced along with the data that is read
and written by the program. Tracing also displays the active DEFINEs, assigns, the
startup message, and process termination messages.
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: