[prev in list] [next in list] [prev in thread] [next in thread] 

List:       pgsql-performance
Subject:    Re: [PERFORM] Curious about dead rows.
From:       "Merlin Moncure" <mmoncure () gmail ! com>
Date:       2007-11-14 3:36:11
Message-ID: b42b73150711131936l65b97b9ya7953383673b2710 () mail ! gmail ! com
[Download RAW message or body]

On Nov 13, 2007 9:26 PM, Jean-David Beyer <jeandavid8@verizon.net> wrote:
> Merlin Moncure wrote:
> > what does pg_stat_all_tables say (assuming row level stats are on)?
> It says stuff like this:
>
>  relname  | seq_scan | seq_tup_read | idx_scan | idx_tup_fetch | n_tup_ins |
> n_tup_upd | n_tup_del
> ----------+----------+--------------+----------+---------------+-----------+-
>  ibd      |       75 |      9503850 |       11 |       2350555 |   2416845 |
>         0 |         0
>  vl_cf    |      139 |     38722575 |       22 |       5392609 |   5692814 |
>         0 |         0
>  vl_li    |      139 |     39992838 |       22 |       5569855 |   5885516 |
>         0 |         0
>
> I removed the relid and schemaname and squeezed the other columns so it
> would not be quite so wide. Is this what you might like to know?

it tells me that you aren't crazy, and that rollbacks are the likely
the cause, although you appear to be watching the logs pretty
carefully.  you can check pg_stat_database to confirm if your
rollbacks are in line with your expectations.  or, you might by seeing
some corner case conditions...are any fields in the table foreign
keyed to another table (cascading update/delete)?  do you have any
functions with handled exceptions or savepoints? (I'm guessing no to
the latter).

merlin

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
       subscribe-nomail command to majordomo@postgresql.org so that your
       message can get through to the mailing list cleanly
[prev in list] [next in list] [prev in thread] [next in thread]