Administration Guide

Setup for AutoLoader

Before running AutoLoader:

1. You must first create a temporary working directory and copy the file /sqllib/samples/autoloader/autoloader.cfg into this temporary directory. This directory is where you should invoke AutoLoader and it has to be NFS-accessible (read and write) to all participating split and load database partitions.
2. Modify the autoloader.cfg file according to the instructions in the sample.
3. The shell program ksh is assumed to be in the /bin directory. If you find it in some other directory, change the first line #!/bin/ksh of all the shell scripts to #!your_ksh_directory/ksh. You can test this using the which ksh or the whence -v ksh command. Also, since the shell scripts use sed, mkdir, awk, cut, grep, egrep, tr, ftp, cat, lsuser, and rsh commands, make sure you have all these programs installed properly. And program mknod is assumed to be in /etc directory. If it is not, please assign a proper value to MKNOD_DIR variable.
4. If data file is remote, put the following information in the $HOME/.netrc file:

      machine machine_name login login_name password login_password
   

   where machine_name is the hostname of the source_system (i.e. the place we need to ftp from to get the data file.

   Ensure that the permissions of .netrc files are -rw------- If not, use the following command to change the permissions.

      $ chmod 600 .netrc
   

5. On each participating database partition (splitting and loading), there must be a not-nfs directory, which has identical path name and is writable. The default is assumed to be /notnfs. A sub-directory with the name $USERID is expected in the notnfs path as well. That is, the path is expected to include: /notnfs/$USERID/.

[ DB2 List of Books | Search the DB2 Books ]