robertgilland wrote:

> According to our DBA a sweep is being run nightly.
> on the database how could then "The gap" be over a million?
>

Two possibilities come to mind. One is that there's an
old transaction lurking when the sweep runs that could
possibly read old versions. The other is that the oldest
transaction isn't rolled back, but stuck in limbo. That
can happen if you use replication, for example.

A big gap between the oldest transaction and the oldest
active was a big deal in the early days of InterBase when
memories were tiny and everyone ran classic. Then, an
array of a few million bits in each process would cripple
a system. Today, I wouldn't worry much about that gap,
except to wonder if sometimes a transaction is left active
longer than expected.

Regards,


Ann