Administration Guide

Troubleshooting

1. If it appears that db2autold is hanging, you can:
   • Check load_log.node_num files to get the progress of load
   • Check load_log.node_num files to get the progress of load on each database partition.
   • Check splt_log.node_num files to see the status of db2split processes on each splitting database partition. If things are going well and the TRACE parameter is set in the db2autold configuration file, there should be trace messages for a certain number of records in these log files.
   • Check db2psplit_msg, db2pread_msg.node_num, and db2pmove_msg.node_num files to see if there are any error messages.
   • If you do find errors that would suggest db2split is not doing anything, or that db2psplit or db2pread or db2pmove encountered errors, then you should interrupt the current db2autold.
2. If there is a failed or interrupted db2autold, the first thing you have to do is to clean up the working environment by issuing:

      db2autold -c your_cfg_file -d
   

   Make sure your_cfg_file is the same configuration file that the failed or interrupted db2autold used.
3. If db2autold still fails you can try the following to diagnose the problem:
   • Change MODE in your db2autold configuration file to SPLIT_ONLY. Run db2autold again. Check the split data files to see if there is anything abnormal in them.
   • If the split files look correct, then try to load one of those split files manually on the right database partition.
   • If the data loads OK, then there might be some db2autold functional problems or database system problems. Please contact your IBM service representative.

[ DB2 List of Books | Search the DB2 Books ]