Re: feature: dynamic DB cache resizing

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> Ed L. wrote:
>> We have ~75 pgsql clusters running in environments where downtime
>> comes at a premium cost. We often run multiple clusters on a
>> single box, and find it necessary to adjust the size of the
>> static DB cache as we add or move clusters. Unfortunately, that
>> means some downtime. It would be extremely useful in minimizing
>> downtime to be able to dynamically resize the DB cache without
>> having to restart the cluster. How hard would that be?

> We can't think of how to do because it is all shared memory shared by
> ever forked process --- does that help explain its complexity?

There isn't any particularly good reason to be resizing shared_buffers
on the fly anyway; much easier to let the kernel adapt the size of its
disk cache instead. Best practice for shared_buffers is to set it
somewhere in the range of 10K to 50K and forget it.

If Ed was thinking of something other than shared_buffers, he should
explain what.

regards, tom lane