Josh Berkus <josh@agliodbs.com> writes:
>> [ shrug ... ] We do not promise that the current partitioning scheme
>> scales to the number of partitions where this is likely to be an
>> interesting concern.
> Actually, you can demonstrate pretty significant response time delays on
> only 50 partitions.
And your point is? The design center for the current setup is maybe 5
or 10 partitions. We didn't intend it to be used for more partitions
than you might have spindles to spread the data across.
>> We're talking "wasted effort on a dead-end situation". The time that
>> would go into this would be much better spent on real partitioning.
> That only applies if someone is working on "real partitioning". Is anyone?
There is discussion going on, and even if there weren't, the argument
still applies. Time spent on this band-aid would be time taken away
from a real solution. In case you haven't noticed, we have very finite
amounts of manpower that's competent to do planner surgery.
regards, tom lane