On Wed, Nov 16, 2016 at 12:58 AM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> It's not really related to lossy pages, it's just that due to deletions
> / insertions a lot more "shapes" of the hashtable are hit.
Okay..
>
> I suspect that this is with parallelism disabled? Without that the query
> ends up using a parallel sequential scan for me.
>
It's with max_parallel_worker_per_gather=2, I always noticed that Q6
takes parallel seq scan only for
max_parallel_worker_per_gather=4 or more..
>
> I've a working fix for this, and for a similar issue Robert found. I'm
> still playing around with it, but basically the fix is to make the
> growth policy a bit more adaptive.
Okay.. Thanks.
--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com