Re: [HACKERS] pgsql 10: hash indexes testing

On Sun, Aug 06, 2017 at 04:32:29PM +1000, AP wrote:
> On Sat, Aug 05, 2017 at 04:41:24PM +0530, Amit Kapila wrote:
> > > (On another note, I committed these patches.)
> > 
> > Thanks.
> 
> Seconded. :)
> 
> Now uploading data with fillfactor of 90. I'll know in 2-3 days
> if the new patches are successful (earlier if they did not help).
> 
> I compiled (as apt.postgresql.org does not provide the latest
> beta3 version for stretch; only sid which has a different perl
> version) 10~beta3~20170805.2225-1~593.git0d1f98b.

Have gotten success with a dataset that failed before with ff 10.

End result:

mdstash=# select * from pgstathashindex('link_datum_id_idx');version | bucket_pages | overflow_pages | bitmap_pages |
unused_pages| live_items | dead_items |   free_percent
 
---------+--------------+----------------+--------------+--------------+------------+------------+------------------
 4 |     12391325 |        5148912 |          158 |       191643 | 4560007478 |          0 | 1894.29056075982
 
(1 row)

mdstash=# select 5148912.0/158/8;      ?column?
-----------------------4073.5063291139240000
(1 row)

The index is 135GB rather than 900GB (from memory/give or take).

I'm happy so far. I'll be seeing how much more I can dump into it this weekend. :)

Thanks muchly to the both of you. :)

AP