I doubt this browsing issue has anything to do with
'random_page_cost'... Why on earth would it? This is a PostgreSQL
setting, to which the ESRI software is essentially completely oblivious.
The ESRI software just sits on top of this PostgreSQL infrastructure, it
essentially doesn't do anythin
ESRI has multiple issues in itself that are terrible to debug.
A client of ours fights a bug where it can't browse tables once you set
random_page_cost to the SSD values of 1.1 instead of 4. So I guess if we're
using a bit more index scan than usual there may be an unknown problem that
does not sh
