We have about a dozen TSM client systems with DB2 running under x86 Linux. These systems use the TSM API interface built into DB2 to back up the DB2 databases. We always use separate node names for DB2 backups and ordinary file backups. About half of the DB2 servers are using the 5.5.1.0 level of the TSM API. Most of these are more or less regularly retrieving archived files during DB2 backups. All of the retrieves come from the node names dedicated to DB2 backups. We have two systems that have done retrieves every day for the last thirty days, four that have done retrieves on some but not all of the last 30 days, and one 5.5.1.0 client that has not done a retrieve within the last 30 days. None of the DB2 servers with older API code have done retrieves during the past 30 days.
OS levels for the DB2 servers performing retrieves are a mixture of 2.4.21-27 and 2.6.18-16. OS levels for the other DB2 servers are a mixture of the same two levels. Our TSM server runs under mainframe Linux. The problem was first noticed with the TSM server code at 5.4.2.0 and has continued after a recent upgrade to 5.5.4.0. Is there a way to stop the retrieves?
