Slow Index Creation, why is it not consuming more memory.

From: Tory M Blue <tmblue(at)gmail(dot)com>
To: "pgsql-performance(at)postgresql(dot)org" <pgsql-performance(at)postgresql(dot)org>
Subject: Slow Index Creation, why is it not consuming more memory.
Date: 2015-12-08 06:36:56
Message-ID: CAEaSS0Zirkg32sea3pUDQSzdxxwQ8DWcNb5v2x2cQiMS-tAKEg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Just trying to figure something out.

9.3.4, CentOS6.5
256GB Ram

Maintenance_work_mem = 125GB
Effective_Cache = 65GB

I have 2 indexes running, started at the same time, they are not small and
one will take 7 hours to complete.

I see almost zero disk access, very minor, not what I want to see when I
have an index that will be as large as it is! , but whatever.

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND

18059 postgres 20 0 5119m 2.4g 20m R 100.0 0.9 12:31.05 postmaster

18091 postgres 20 0 9794m 6.9g 20m R 100.0 2.7 11:41.61 postmaster
Why are 2 index using different amounts of resident mem, when they have the
keys to the castle, and why are they not taking more?

I've tried this with as low as 4GB of maintenance_work_mem and the numbers
in TOP stay the same.

Work_mem is 7GB.

What am I not understanding missing?

Thanks
Tory

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Peter Geoghegan 2015-12-08 06:40:21 Re: Slow Index Creation, why is it not consuming more memory.
Previous Message Brendan McCollam 2015-12-04 19:34:29 ossp-uuid: Performance considerations for different UUID approaches?