Hitachi

Hitachi Advanced Database Command Reference


33.6 Notes

  1. The adbmergechunk command uses SQL statements to retrieve table data that has been stored. For this reason, we recommend that you execute the command in an environment where out-of-order execution is applied. For details about environments in which out-of-order execution is applied, see Database retrieval method (out-of-order execution) in the HADB Setup and Operation Guide.

  2. Use X'0A' (LF) or X'0D0A' (CRLF) for the line break code in any file specified in the options of the adbmergechunk command.

  3. Use the adbcancel command to cancel execution of the adbmergechunk command. Alternatively, use a method such as the adbstop --force command or the OS's kill command to forcibly terminate the server process. After you have canceled the adbmergechunk command, see 33.5 Handling abnormal termination of the adbmergechunk command and take appropriate action as needed.

    For details about how to use the adbcancel command to suspend an executing command, see 1.6 When suspending (forcibly terminating) an executing command.

  4. Merge chunk processing is performed by server processes. Therefore, using the Ctrl+C keys to forcibly terminate the adbmergechunk command process will not interrupt merge chunk processing that is being performed by server processes (such processing continues). When the merge chunk processing is completed, the command outputs the KFAA81001-I message to the message log file.

  5. Work files (temporary work files) are created during merge chunk processing. These temporary work files are deleted automatically when the adbmergechunk command terminates normally (the return code is 0 or 4). If the adbmergechunk command does not terminate normally, these work files might not be deleted.

    Some of the temporary work files are needed when other commands are re-executed. For this reason, do not delete temporary work files before you re-execute other commands. If you delete them, the other commands will not execute correctly.

  6. Because temporary work files might cause a shortage of disk space, we recommend that you delete temporary work files from time to time if there is no need to re-execute other commands.

    Check the following directories periodically for any remaining temporary work files and delete any that are found:

    • $DBDIR/ADBSYS/ADBUTL

    • Temporary work file storage directories specified in the -w option (or in $DBDIR/ADBWORK when the -w option was omitted)

  7. You cannot execute the adbmergechunk command on the following tables:

    • A table for which any indexes that are in unfinished status are defined

    • A non-updatable table

    • A table for which execution of the adbimport or adbidxrebuild command has been canceled

  8. Make sure that there is no duplication of the absolute paths for the directories and files specified in the adbmergechunk command's options. This applies to the following directories and files:

    • Temporary work file storage directories (specified in the -w option or specified in the directory path file specified in the -w option)

    • Directory path files specifying the storage locations of temporary work files (specified in the -w option)

    • Merge chunk option file (specified in the -z option)

    Make sure that the absolute paths of these directories and files are unique from those of any file names used by any programs other than the adbmergechunk command. If a duplicate file name is used, data in the file might be corrupted. Also make sure that symbolic links and relative paths are specified correctly.

  9. If the adbmergechunk command is executed while another command or application program is running, HADB might not be able to allocate the required number of processing real threads. If this happens, HADB waits until the required number of processing real threads can be allocated before executing the adbmergechunk command. If allocation of processing real threads is placed in wait status, see the topic Points to consider about the number of processing real threads to be used during command execution in Points to consider when executing commands concurrently in Preparing Resources in the HADB Setup and Operation Guide and take appropriate action.

  10. To reduce this command's execution time, see the topic Reducing the execution time of the adbmergechunk command in Tuning to shorten command execution time in Tuning in the HADB Setup and Operation Guide.