Tim Uckun <timuckun@gmail.com> writes:
>> Hard to tell. We've seen enough reports like that to make it seem like
>> there may be some bug buried there, but no one has provided anything to
>> do any debugging work with. Can you create a reproducible test case?
> Not really. I have a nightly process which downloads data and sticks
> it into a text field. Afterwards another process reads that text data
> and processes it creating rows in another table. The problem occurs in
> the last step and at seemingly random intervals. For example one time
> it might happen when you are creating row 1000 another time it might
> be when you are creating row 2000.
Well, I'm not asking for perfect reproducibility --- a test case that
fails even 1% of the time would be great.
regards, tom lane