IBM Books

Command Reference

db2autold - Autoloader

Autoloader is a tool for splitting and loading data in an MPP environment. This utility can:

Autoloader can run in one of four modes:

Split_And_Load
In this mode, data is FTPed from the input source (if it is remote). It is then split in parallel and loaded simultaneouly on the corresponding nodes.
Split_Only
In this mode, data is FTPed from the input source (if it is remote). It is then split in parallel, and the output from the splitters is written to files in the current Autoloader working directory.
Load_Only
In this mode, data is assumed to be pre-split and contained in files named as follows: filename.xxx, where xxx represents the node number. The split process is skipped, and data is loaded simultaneously on the corresponding nodes. It is assumed that the files reside in the current Autoloader working directory.
Analyze
In this mode, an optimal partitioning map with even distribution across all nodes is generated.
Note:This utility does not split LOBs.

Authorization

sysadm

Required Connection

None. This command establishes a database connection.

Command Syntax



>>-db2autold---+-----+--+-----+--+-----+--+-----+--------------->
               +--h--+  +--d--+  +--D--+  +--i--+
 
>--+------------------+----------------------------------------><
   +--c--config_file--+
 

Command Parameters

-h
Display help information. When this option is specified, all other options are ignored, and only the help information is displayed.
-d
In the case of an abnormal exit from Autoloader, it is necessary to run Autoloader with this option to clean up all associated temporary directories, files, and hanging processes.
-D
Turn on debug mode.
-i
Turn on interactive mode for clean-up.
-c config_file
Name of the file containing configuration data required by the tool. The default name is autoload.cfg.

Usage Notes

Autoloader creates a log file called autoload.logfile. This file contains messages from the main autoloader script, and should be checked to ensure that all pipes and temporary directories have been set up correctly. The utility also creates a file called load_log.xxx. This file contains messages from the load process on node xx. The utility also creates a file called splt_log.xxx. This file contains messages from the split process on node xx.

If the order of input data is to be maintained during a load operation, only one node should be used for splitting the data. Parallel splitting does not guarantee that the data will be loaded in the order that it was received.

If LOBs are stored in separate files (that is, the LOAD LOBSINFILE file type modifier is being used), all directories containing the LOB files should be made accessible to all the nodes involved in the load operation.

Autoloader ignores the LOAD messages parameter, and directs all messages from the LOAD command into load_log.xxx.

If the STATISTICS YES option is specified with the LOAD command, Autoloader chooses only one output node on which to collect statistics.

For more information about db2autold, see the Administration Guide.

See Also

db2split - Data Declustering Tool


[ Top of Page | Previous Page | Next Page | Table of Contents | Index ]

[ DB2 List of Books | Search the DB2 Books ]