The customer I am working with has been struggling with RDRS v6.1 in a z/VSE environment for well over a year without being able to actually get the product into production for various reasons--though many issues are being resolved successfully. The biggest remaining reason is the performance (or lack thereof) for RDRS CDC in capturing batch VSAM changes and sending them down to DB2/LUW on the network side.
Initially we were told that the problem was slow TCP/IP transmissions. We resoled that with a 900+% improvement. But my observations, while monitoring large batch update jobs, are that CDC isn’t even *starting* to download accumulated VSAM changes until up to 5 minutes *after* the update job has ended! Thus, even if that batch job didn’t overflow the 1G total dataspace available for the pool, follow-on batch update jobs would--resulting in “TVS066E Data Lost”.
Then, we were finally told that RDRS v6.1 is not intended to keep up with the rate at which we are generating VSAM changes in batch. What!?! Further, that we would have to write our own “Batch Compare and Synchronize” process to apply these changes, ourselves, after-the-fact. What!?! Surely we are not such a unique environment that we might be the “only” ones who are having trouble with RDRS CDC performance in a VSE environment.
So… Any other VSE customers out there willing to discuss how to improve the performance of RDRS v6.1 CDC in a batch VSE environment?
