On 06/27/2015 12:30 AM, Jim Nasby wrote:
> On 6/24/15 6:41 PM, Tomas Vondra wrote:
>>> Were the stories (or the experience which lead to the stories) on
>>> 9.3 or later? Do they have a good way to reproduce it for testing
>>> purposes?
>>
>> The per-db split can only improve things if there actually are
>> multiple databases, and if the objects are somehow evenly
>> distributed among them. If there's a single database, or if most of
>> the objects are in a single database (but stored in multiple
>> schemas, for example), it does not help much. So it's trivially to
>> reproduce the previous issues.
>
> Right, and a single large database is a pretty common scenario.
FWIW I doubt this scenario is so common.
Most of the cases of issues with stats file I've heard of were about a
database server shared by many applications, separated into databases.
That's basically the scenario we've been facing with our PostgreSQL
machines, and the motivation for the split.
Maybe there are many clusters with a single database, containing many
objects, but I don't remember any recent reports of problems with stats
files from them. Those pretty much disappeared after 9.3, which is when
the split happened.
That is not to say we can just design it in a way that will cause OOM
issues or crashes on such machines ...
regards
--
Tomas Vondra http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services