Date: Fri, 19 Dec 2008 10:04:46 -0300
From: Alvaro Herrera <alvherre@commandprompt.com>
To: Laszlo Nagy <gandalf@shopzeus.com>
Cc: pgsql-admin@postgresql.org
Subject: Re: rebellious pg stats collector (reopened case)
Message-ID: <20081219130446.GC4278@alvh.no-ip.org>
References: <49479876.8040607@shopzeus.com>
	<20081216134542.GH4741@alvh.no-ip.org>
	<494B614C.5060808@shopzeus.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <494B614C.5060808@shopzeus.com>
User-Agent: Mutt/1.5.18 (2008-05-17)

Laszlo Nagy wrote:
> Alvaro Herrera wrote:
>> Laszlo Nagy wrote:
>>   
>>> PostgreSQL 8.3.5, the system is now stable (uptime > 10 days).   
>>> PostgreSQL stats collector uses 100% CPU forever:
>>>     
>>
>> Could you grab a few stack traces from it and post them?  Also possibly
>> useful, leave strace running on the pgstat process for a while and post
>> the output somewhere.
>>   
> Very interesting results. Before I used 'truss' on the process, it was:
>
> 78816 pgsql            1 106    0 22400K  7100K CPU6   6  24.2H 99.02%  
> postgres
>
> After I started "truss -p 78815" the first message I got was:

It was 78816 and you traced 78815?  Are you sure the process with 24h of
CPU was pgstat?

-- 
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.