Re: Change in Log Format and Prepared Statements

From: Alexander Stanier <alexander(dot)stanier(at)egsgroup(dot)com>
To: Craig Ringer <ringerc(at)ringerc(dot)id(dot)au>
Cc: Dave Cramer <pg(at)fastcrypt(dot)com>, pgsql-jdbc(at)postgresql(dot)org
Subject: Re: Change in Log Format and Prepared Statements
Date: 2012-09-17 06:45:14
Message-ID: 5056C6FA.1020800@egsgroup.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-jdbc

<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<font size="-1"><font face="Helvetica, Arial, sans-serif">Thanks
Dave and Craig for you input.<br>
<br>
Does anyone know what the three durations specifically
represent? Presumably two of them are the time to prepare and
execute the statement? What's the third? Is the overall
execution time the sum of the three (or two in some cases)?<br>
<br>
It would be useful to have one summary time for the overall
execution of each statement as before. We wrote a little Java
program which analyses log file based on a statement followed by
single duration (and outputs the most frequent, the slowest and
most costly queries). We could rewrite the program to cope with
the new format if needs be but if we can get a single duration
it would save us some work. I don't mind using prepared
statements if there really is no overhead and I can fully
understand what each part of the new log format is.<br>
<br>
Alex Stanier.<br>
<br>
</font></font>
<div class="moz-cite-prefix">On 17/09/2012 06:15, Craig Ringer
wrote:<br>
</div>
<blockquote cite="mid:5056B201(dot)5020809(at)ringerc(dot)id(dot)au" type="cite">On
09/14/2012 03:24 AM, Dave Cramer wrote:
<br>
<blockquote type="cite">Yes, the new driver uses V3 protocol which
uses prepared statements,
<br>
named and un-named. I think the 3 durations may be due to the
prepare
<br>
and execute.
<br>
</blockquote>
<br>
I noticed this as well, and wasn't entirely thrilled with the
effects on logging.
<br>
<br>
I wonder if Pg can be convinced to suppress the non-execution
times, or merge them to report a single statement execution time.
I'll poke -hackers.
<br>
<br>
--
<br>
Craig Ringer
<br>
<br>
<br>
<br>
This email has been scanned for viruses by Blackspider's Mail
Control System. For more information please visit
<a class="moz-txt-link-freetext" href="http://www.blackspider.com">http://www.blackspider.com</a>&nbsp; </blockquote>
<br>
<br><br>
<P align=center><FONT style="BACKGROUND-COLOR: #ffffff">This message has been scanned for malware by SurfControl plc. </FONT><A href="http://www.surfcontrol.com/"><FONT style="BACKGROUND-COLOR: #ffffff" color=#000000>www.surfcontrol.com</FONT></A></P>
</body>
</html>

Attachment Content-Type Size
unknown_filename text/html 2.6 KB

In response to

Responses

Browse pgsql-jdbc by date

  From Date Subject
Next Message Alexander Stanier 2012-09-17 06:58:21 Re: Change in Log Format and Prepared Statements
Previous Message Craig Ringer 2012-09-17 05:15:45 Re: Change in Log Format and Prepared Statements