G
Garry Hall
Guest
Without knowing all the details, I would tend to agree with my colleagues in Tech Support. This sounds more like memory corruption. A memory leak alone would probably not corrupt the data, it would just consume the available memory, but (typically) the content of that leaked memory would be intact. In the bug attached to the KB you mentioned, if you walk the FIRST-QUERY chain, the error message indicates the leaked queries know they were auto-generated for a dataset. Of course, this bug has not been investigated, so what might look to be a leak might actually be corruption (the "right" bit just happens to be turned on to give this message). It seems unlikely, though, if the sanity of the chain is sufficient to be able to walk it. If this is memory corruption, I think there is little that can be done about it now, other than cleaning up. Memory corruption is hard to track down at the best of times, and requires a reproducible case and/or a very dedicated user willing to work with TS/Development.
Continue reading...
Continue reading...