From: Aldor <an@mediaroot.de>
To: Michael Fuhr <mike@fuhr.org>
Cc: Marc G. Fournier <scrappy@postgresql.org>
Cc: pgsql-admin@postgresql.org
Subject: Re: What is syslog:duration reporting ... ?
Date: Fri, 26 Aug 2005 00:50:29 +0100
Message-ID: <430E5945.4000408@mediaroot.de> (raw)
In-Reply-To: <20050825195940.GA54140@winnie.fuhr.org>
References: <20050825152428.T1044@ganymede.hub.org>
<430E12F1.1080206@mediaroot.de>
<20050825155331.V1044@ganymede.hub.org>
<20050825195508.GA54074@winnie.fuhr.org>
<20050825195940.GA54140@winnie.fuhr.org>
Aug 25 14:53:32 forgehouse-s1 postgres[23721]: [2-1] LOG: duration:
567.559 ms statement: UPDATE session SET hit_time=now() WHERE
md5='7537b74eab488de54d6e0167d1919207';
Aug 25 14:53:32 forgehouse-s1 postgres[23722]: [2-1] LOG: duration:
565.966 ms statement: UPDATE session SET hit_time=now() WHERE
md5='d84613009a95296fb511c2cb051ad618';
Aug 25 14:53:33 forgehouse-s1 postgres[23723]: [2-1] LOG: duration:
1192.789 ms statement: UPDATE session SET hit_time=now() WHERE
md5='d84613009a95296fb511c2cb051ad618';
Aug 25 14:53:53 forgehouse-s1 postgres[23727]: [2-1] LOG: duration:
12159.162 ms statement: UPDATE session SET hit_time=now() WHERE
md5='d84613009a95296fb511c2cb051ad618';
Aug 25 14:53:54 forgehouse-s1 postgres[23728]: [2-1] LOG: duration:
3283.185 ms statement: UPDATE session SET hit_time=now() WHERE
md5='7537b74eab488de54d6e0167d1919207';
Aug 25 14:53:57 forgehouse-s1 postgres[23729]: [2-1] LOG: duration:
2116.516 ms statement: UPDATE session SET hit_time=now() WHERE
md5='7537b74eab488de54d6e0167d1919207';
Take a look to the timestamps... they are not really close to each other...
Michael Fuhr wrote:
> On Thu, Aug 25, 2005 at 01:55:08PM -0600, Michael Fuhr wrote:
>
>>Are the updates happening inside a transaction? Is it possible
>>that they're being blocked by other transactions that update the
>>same record around the same time and don't commit immediately?
>>I can duplicate the results you're seeing by doing that.
>
>
> As for why you don't see long durations with EXPLAIN ANALYZE, here's
> a possible explanation: "ordinary" updates of the same record might
> happen close together because that's how the application works, but
> by the time you run EXPLAIN ANALYZE no other transactions are updating
> that record so the update executes immediately.
>
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients from your mail client; the link fills in
the headers below: reply via email
Some clients, webmail especially, drop In-Reply-To and References from
a link, and the reply then starts a new thread: use the mbox below.
To: pgsql-admin@postgresql.org
Cc: an@mediaroot.de, mike@fuhr.org, scrappy@postgresql.org
Subject: Re: What is syslog:duration reporting ... ?
In-Reply-To: <430E5945.4000408@mediaroot.de>
References: <20050825152428.T1044@ganymede.hub.org> <430E12F1.1080206@mediaroot.de> <20050825155331.V1044@ganymede.hub.org> <20050825195508.GA54074@winnie.fuhr.org> <20050825195940.GA54140@winnie.fuhr.org> <430E5945.4000408@mediaroot.de>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox