DUMPANALYZER Files

DUMPANALYZER uses several files, including OPTIONS, TAPEIN, MCPCODEFILE, COREMAP, and DP, among others. You can File-equate these files if you use a Command and Edit (CANDE) or Work Flow Language (WFL) RUN statement or Menu-Assisted Resource Control (MARC) to run DUMPANALYZER. The descriptions of these files follow.

OPTIONS

OPTIONS is the file name of the user input file. To enable you to enter DUMPANALYZER commands at the ODT, you can either enter the DA (Dump Analyzer) system command at the ODT, or you can file-equate OPTIONS to ODT. “ODT Operation” later in this section explains the methods for starting DUMPANALYZER from an ODT. You can also file-equate OPTIONS to a card file containing DUMPANALYZER commands.

TAPEIN

DUMPANALYZER reads the memory dump or program dump that is to be analyzed with the TAPEIN file. When you run DUMPANALYZER, your method for naming the dump file will depend on which of the following circumstances applies:

  • If the dump file is a memory dump on tape, the default applies. The default for TAPEIN is

    FILES TAPEIN(KIND=TAPE,NEWFILE=FALSE,FILENAME=MEMORY/DUMP);
  • If the dump file is a memory dump on disk, you file-equate TAPEIN to the name of the memory dump disk file. A memory dump file on disk has a title in the following format:

    DP/<MMDDYY>/<HHMMSS>/<reason>. 

    Refer to the DN (Dump Name) and the DF (Empty Dumpdisk File) system commands in the System Commands Operations Reference Manual for details about how to create a memory dump on disk.

  • If the dump file has been saved by the SAVE command, see “Saved Memory Dumps” later in this section.

  • If the dump file is a program dump on disk, you file-equate TAPEIN to the name of the program dump disk file. A program dump on disk has a title in the following format:

    PDUMP/<code file title>/<YYMMDD>/<HHMMSS>/<mix number>
  • If the memory dump file is a stream file on disk, you file-equate TAPEIN to the name opf the stream file. One example of a stream file is an MCP Environment dump image taken on a Virtual Machine for ClearPath MCP server. The file name is usually MCPDUMP.ZDP.

MCPCODEFILE

You need to file-equate the MCPCODEFILE file only when the dumping MCP is no longer available on the system performing the analysis. For example, this situation occurs if you remove the dumping MCP code file before DUMPANALYZER begins the analysis of the dump. To avoid this problem, refer to “Saved Memory Dumps” later in this section.

If, for any reason, the dumping MCP is not available when the analysis of the dump begins, you can substitute a similar MCP by label-equating MCPCODEFILE to an MCP codefile that closely matches the dumping MCP.

For example, suppose that the dumping MCP is version 40.100.1234. A suitable substitute might be version 40.101.1256, but version 39.123.2345 would be too different to be suitable.

Note: Label-equating MCPCODEFILE to an MCP other than the dumping MCP might cause unpredictable results, no matter how similar the two MCP code files are to each other.

To warn you that the dumping MCP and the file-equated MCP differ from each other, DUMPANALYZER issues the following message:

WRONG CODEFILE

It then waits for your response to the following ACCEPT statement before continuing:

ACCEPT:  AX OK TO USE <file-equated MCP title> OR
         AX RESTART TO SPECIFY NEW TITLE

You do not need to file-equate MCPCODEFILE if you are analyzing a previously saved dump. For this reason, it is recommended that you create a saved dump immediately after a dump finishes, even if you plan to do the analysis at a later time.

DUMPANALYZER retains the name of the MCPCODEFILE in the pseudorecovery file. Therefore, you do not need to file-equate the MCPCODEFILE if you are running with a pseudorecovery file that contained an MCP file-equate from the previous session. Note that you can override the saved MCPCODEFILE name by file-equating a new file when you restart the analysis with the pseudorecovery file.

COREMAP

You can file-equate the COREMAP file if you want to analyze a COREMAP data file without its corresponding memory dump. For example,

FILE COREMAP=MEMDUMP/COREMAP/112108/1700 ON DUMPS 

When you run DUMPANALYZER using this method, DUMPANALYZER only recognizes a limited number of commands, such as the COREMAP command and commands that control output.

DP (Pseudorecovery File)

While DUMPANALYZER is initializing, it accepts the MEMORY/DUMP file as input and builds certain data structures. When it completes the initialization process, DUMPANALYZER saves the data structure information in a pseudorecovery file under your usercode. This file has the name

DP/<MMDDYY>/<HHMMSS>/<REASON_ID>

MMDDYY represents month, day, and year, and HHMMSS represents hour, minutes, and seconds.

You can override the title and family of the pseudorecovery file for the DP file by file-equating its TITLE, FILENAME and/or FAMILYNAME. This allows you to direct the file to a family other than your default family and name the pseudorecovery file with a user-defined name. Note that the DP file-equation is ignored if the input file (TAPEIN) is a disk file.

If the DUMPANALYZER run is interrupted by a halt/load or terminated by an operator, you can file-equate TAPEIN to this pseudorecovery file and use it as input to DUMPANALYZER. When you use a pseudorecovery file, you must file-equate the MCPCODEFILE to the MCP code file that was in use at the time of the dump.

If for any reason you want to exit from the DUMPANALYZER session and restart it again later, you can save time if you use the RECESS command rather than the STOP or BYE command. The RECESS command does not remove the pseudorecovery file, whereas the other two commands do. In all cases, DUMPANALYZER does not remove the file that was used as input.

DUMPANALYZER removes the pseudorecovery file if the SAVE command is issued and successfully completed. For advice on when to use the SAVE command, refer to “Saved Memory Dumps” later in this section.

DUMPANALYZER can request exclusive access to the pseudorecovery file during initialization or while it is processing a SAVE command. This exclusive access is necessary to prevent any interference with other instances of DUMPANALYZER that are using the same pseudorecovery file. If DUMPANALYZER does have exclusive access to the pseudorecovery file and another DUMPANALYZER transaction is initiated, one of the processes must wait, and the following message appears:

WAITING FOR:<pseudorecovery file title>

If the delay is not acceptable, you can start the second instance with a different primary family for DISK, use a different usercode, or file-equate the DP file using a different TITLE, FILENAME or FAMILYNAME.