The dump files are located in the directory filebase/DUMP or filebase\dump, where filebase is the base name of the KDCFILE in the MAX statement.
If the dump is caused by an error in the UTM system code (system dump), a file generation group (FGG) is created (see below).
Otherwise, the name of the dump file has the following format:
reason.pid.no
where
| reason | ID identifying the cause of the memory dump (6 characters). |
| pid | Process ID number (pid) under which the dump was produced. |
| no | Running number of the dumps generated by a process in an application if the dump files are not created as a file generation group (FGG); or hexadecimal value of the counter for the number cold starts of the application if the dump files were created as FGG files. |
In a system dump, a file generation group (FGG) is set up in the directory reason.pid.no. If two or more work processes are running in the application, one dump file is set up for each process. The FGG has the form:
/INFO | (FGG administration file) |
/0001 | (first dump file) |
/0002 | (second dump file, etc.; cf. USLOG-FGG) |
. |
If the UTM application was started with TESTMODE=OFF, and if a PEND ER occurs followed by one of the following KCRCDC codes, the UTM dump is suppressed.
FH01, K301, K302, K345, K601, K602, K603, K608, KM01, KM02, KM03, KM04, KM05, KM07, KM08, KR01, KR02, KT01, KT02, KT04, KU14
In this event, the work process is not terminated or restarted, i.e. the work process continues working for the UTM application. Exception: PGWT calls are permitted for the current TAC, and the call at which the error occurred was not a PEND call.
In the case of the KCRCDC code K316, no UTM dump is written, regardless of test mode. If PGWT calls are permitted for the current TAC, and the call at which the error occurred was not a PEND call, the program is loaded.
Reducing the volume of dump information with the start parameter DUMP-CONTENT
The start parameter DUMP-CONTENT allows you to specify whether openUTM is to reduce the volume of dump information or not. In this case, reduction means that process-independent memory areas (shared memories) are only included in the dump of the work process which caused the application to abort. Reducing the dump information means that the diagnostic documentation in the event of abortion of an application requires far less space. Reduction of the dump information is activated by default. The start parameter DUMP-CONTENT can be used to deactivate or reactivate reduction of the dump information as required.
Syntax of the start parameter:
.UTM START DUMP-CONTENT={ STANDARD | EXTENDED }
STANDARD | When UTM creates a dump file generation, process-independent memory areas are only contained in the dump for the first process (which caused abortion). This is generally sufficient for diagnostic purposes and is set by default. |
EXTENDED | Process-independent memory areas are contained in all the dumps of a dump file generation. You should only set this value when required to explicitly by the software support. |