Horton
Menu

Re: Should pg 11 use a lot more memory building an spgist index?

Bruno Wolff III <bruno@wolff.to>

View thread
From: Bruno Wolff III <bruno@wolff.to>
To: pgsql-general@lists.postgresql.org
Date: 2018-10-25T12:10:29Z
Lists: pgsql-hackers 
It looks like it got past creating the exclude constraint based on the 
ordering of commands in the dump file. However creating a more normal 
spgist index is taking a very long time with a lot of disk wait time. 
CPU usage seems pretty low for the amount of time it has been working 
on building that index, but maybe that is normal for building indexes.
I used the almost the default postgresql.conf on my workstation. I bumped 
up the limits in a few places on the server that could have allowed a lot 
more memory to be used especially if the index creation was parallelized. 
While the load is running I'll see if I can tell if there is a memory leak 
with this index build. Once it finishes, I can dump a specific table 
and test building the exclude spgist index with some different settings 
to see if I can reproduce the out of memory error with a lot less data 
then is in the whole database.

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix memory leak in repeated SPGIST index scans.

  2. Add support for nearest-neighbor (KNN) searches to SP-GiST