Re: atrocious update performance

From: "Aaron Werman" <awerman2(at)hotmail(dot)com>
To: "Rosser Schwarz" <rschwarz(at)totalcardinc(dot)com>
Cc: <pgsql-performance(at)postgresql(dot)org>
Subject: Re: atrocious update performance
Date: 2004-03-17 03:25:16
Message-ID: Law10-OE28zi95HfqR7000142f8@hotmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Quick observations:

1. We have an explanation for what's going on, based on the message being
exactly 666 lines long :-)
2. I'm clueless on the output, but perhaps Tom can see something. A quick
glance shows that the strace seemed to run 27 seconds, during which it did:
count| call
-------|---------
84 | _llseek
40 | brk
54 | close
88 | open
63 | read
in other words, nothing much (though it did *a lot* of opens and closes of
db files to do nothing ).

Can you do another strace for a few minutes against the actual update query
adding the -c/-t options and control-c out?

----- Original Message -----
From: "Rosser Schwarz" <rschwarz(at)totalcardinc(dot)com>
To: "'Tom Lane'" <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: <pgsql-performance(at)postgresql(dot)org>
Sent: Tuesday, March 16, 2004 7:20 PM
Subject: Re: [PERFORM] atrocious update performance

while you weren't looking, Tom Lane wrote:

[trace]

`strace -p 21882` run behind the below query and plan ... below that.

# explain update account.cust set prodid = tempprod.prodid, subprodid =
tempprod.subprodid where origid = tempprod.debtid;

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Joe Conway 2004-03-17 04:49:01 Re: rapid degradation after postmaster restart
Previous Message Tom Lane 2004-03-17 01:22:04 Re: atrocious update performance