We have a 4.2.3.2 TSM server running under OS/390. We expect to add more TSM clients, and the workloads on two other mainframe applicatios are increasing. The mainframe is already short of real memory, and is heading rapidly for shortages of CPU capacity and network bandwidth. One of the options being discussed is moving the TSM server to AIX. At least one manager has argued that we should consider migrating to a different enterprise backup solution. One of his arguments is that we are getting poor quality code and poor support.
We have been struggling with one or more memory leaks for a couple of months. Patch level upgrades have improved performance somewhat, but have not fixed the problem. We are still finding it necessary to restart the TSM server twice a day to maintain acceptable performance. Back when we decided to upgrade to the 4.2.3.2 level I went to the anonymous FTP server and found a text file referring me to the TSM server README for information on a password protected FTP server. The information was not in the README file, so I sent out a query on this list and opened a severity 2 incident with IBM/Tivoli. I started getting responses from the list in a couple of hours. IBM/Tivoli finally saw fit to contact me after five days. We have since opened another serverity 2 incident for the apparent memory leak. It has typically taken days of repeated calls to get administrative responses, such as confirmation that dumps were received. It took weeks to get a substantive technical response. We were recently told that our symptoms match open APAR PQ69840. I got a copy of the APAR description. The alleged match is so poor that I am struggling to convince myself that the claim was a mistake and not a deliberate lie. I welcome any suggestions for rebutting the argument that it is time to look for a product with better code and better support.
