Re: issue with gininsert under very high load

* Andres Freund (andres@2ndquadrant.com) wrote:
> There's previously been talk about changing the limits to something more
> reasonable but it got stalled in bikeshedding IIRC.

As I recall, there was argument that we didn't really need a new GUC for
this (which was the proposal) but rather just need to pick a reasonable
(small) value and hard-code it.  Are there objections to doing so, or
are there cases where that would be a serious problem?  How do people
feel about 4MB?
Thanks,
    Stephen