Boris wrote:
> Another interesting point is that, when I paste the slow query into
> psql, and explain/analyze it, bad plan is chosen (like observed from
> the code with new driver).
>
> Any ideas? I'd be happy to provide more details if needed. Is the
> issue somehow caused by the fact that 8.4 driver uses real server side
> prepared statements vs the old one that didn't?
If you get the same problem via psql, I doubt it's related to your
driver changes. What changed on the server side?
-O
