Re: Nooby Q: Should this take five hours? And counting?

From: Gregory Stark <stark(at)enterprisedb(dot)com>
To: Kenneth Tilton <kentilton(at)gmail(dot)com>
Cc: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>, pgsql-general(at)postgresql(dot)org
Subject: Re: Nooby Q: Should this take five hours? And counting?
Date: 2009-04-20 10:32:25
Message-ID: 87k55fwr7q.fsf@oxford.xeocode.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Kenneth Tilton <kentilton(at)gmail(dot)com> writes:

> A bit. I killed the indexing and jacked work_mem up to 500mb, indexing then
> finished in 7 min 25s.
>
> Yer a genius!

FWIW creating indexes using maintenance_work_mem. I would not expect changing
work_mem to really help much. That's, uh, curious.

>> 2: You've got a slow disk subsystem, if you're already seeing 25%
>> IOWait with only ~2 to 3 megs a second being written.
>
> This has been passed along to management for consideration.

Depends. If it's all random i/o then 8-12MB/s is about right for 4-6 drives.
If there's any sequential i/o mixed in then yeah, it's pretty poor.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
Ask me about EnterpriseDB's On-Demand Production Tuning

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Robson Fidalgo 2009-04-20 11:13:15 Re: [DOCS] Postgresql 8.3X supports Arrays of Composite Types?
Previous Message Gregory Stark 2009-04-20 10:29:16 Re: DISTINCT ON without ORDER BY