| From: | Satoshi Nagayasu <nagayasus(at)nttdata(dot)co(dot)jp> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | "Marc G(dot) Fournier" <scrappy(at)postgresql(dot)org>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: sort_mem statistics ... |
| Date: | 2005-10-19 00:33:32 |
| Message-ID: | 4355945C.6030909@nttdata.co.jp |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Tom,
Tom Lane wrote:
>>do we maintain anything anywhere for this? mainly, some way of
>>determining # of 'sorts to disk' vs 'sort in memory', to determine whether
>>or not sort_mem is set to a good value?
>
> As of 8.1 you could turn on trace_sort to collect some data about this.
Why is the trace_sort option DEVELOPER_OPTIONS?
I think the sort statistics are *very* important for DBAs,
not only for developers (hackers).
Without any numerical evidence, trying (and error) to fitwork_mem value
will be painfull and wasting DBA's time.
And I want to get statistic info through system views, like pg_statio_*.
Please remember my previous post.
http://archives.postgresql.org/pgsql-patches/2005-09/msg00116.php
--
NAGAYASU Satoshi <nagayasus(at)nttdata(dot)co(dot)jp>
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Josh Berkus | 2005-10-19 00:55:34 | Re: sort_mem statistics ... |
| Previous Message | Josh Berkus | 2005-10-18 23:26:06 | Re: Seeing context switch storm with 10/13 snapshot of |