From: | Alan Hodgson <ahodgson(at)simkin(dot)ca> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Puzzling full database lock |
Date: | 2012-02-02 00:55:35 |
Message-ID: | 1856771.dLdDo9FaAs@skynet.simkin.ca |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Wed, Feb 1, 2012 at 7:38 PM, Christopher Opena <counterveil(at)gmail(dot)com>
wrote:
> > Hello folks,
> >
> > We've been running into some very strange issues of late with our
> > PostgreSQL database(s). We have an issue where a couple of queries
> > push high CPU on a few of our processors and the entire database locks
> > (reads, writes, console cannot be achieved unless the high CPU query
> > procs are killed). Further investigation shows ~59% total cpu usage
> > (we have 16 total cores), low io, and mid-to-low memory usage (we have
> > 74GB of memory, shared_buffers=16GB).
Define "low I/O". The only things I've ever seen interfere with console access
are running out of memory or excessive disk i/o. If you're seeing even 6-12%
iowait on a 16 core machine you might very well have swamped your available
disk I/O.
From | Date | Subject | |
---|---|---|---|
Next Message | Scott Frankel | 2012-02-02 00:55:57 | Re: restart server on Lion |
Previous Message | Carlos Mennens | 2012-02-02 00:55:30 | Re: Puzzling full database lock |