Message-ID: <49218C90.40804@egsgroup.com>
Date: Mon, 17 Nov 2008 15:24:00 +0000
From: Ben Webber <ben.webber@egsgroup.com>
User-Agent: Thunderbird 2.0.0.17 (X11/20081109)
MIME-Version: 1.0
To: Alvaro Herrera <alvherre@commandprompt.com>
Cc: pgsql-admin@postgresql.org,
	Alexander Stanier <alexander.stanier@egsgroup.com>
Subject: Re: Change in logging behaviour between 8.1 and 8.2
References: <4901E4ED.7090301@egsgroup.com> <6092.1224892561@sss.pgh.pa.us>
	<49217587.4030709@egsgroup.com>
	<20081117145837.GG4291@alvh.no-ip.org>
In-Reply-To: <20081117145837.GG4291@alvh.no-ip.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Sorry, meant 2 gigs, not 10.

An interesting suggestion, but the problem with storing the logfiles in
a table for us would be that uncompressed, the log file for a day is
about 10 gigs. This would mean that an unacceptable amount of excess
data would accumulate in the database. It would be feasible however to
write a script to import the archived logfile into a new temporary
database on a different server, then use SQL to search it and delete the
db when finished.

Thanks for the suggestion though.

Ben.


Alvaro Herrera wrote:
> Ben Webber wrote:
>> Hi,
>>
>> I wrote a shell script to find the duration and the related statement in the log file and place them one after the other if the duration is over a specified time like this:-
>>
>> 2008-10-31 02:00:49 GMT [23683] [mp_live] LOG:  statement: CLUSTER;
>> 2008-10-31 02:04:42 GMT [23683] [mp_live] LOG: duration: 232783.684 ms
> 
> I wonder if you'd benefit from doing CSV logs and then storing them into
> a table.  Querying using SQL is probably going to be easier (and more
> robust -- it'd work even with embedded newlines etc).
> 


This message has been scanned for malware by SurfControl plc. www.surfcontrol.com