Обсуждение: Can not deleted all record

Yudha,

"ERROR:  Query was cancelled." usually means that the query is taking too long,
and you exceeded the maximum time specified in your postgresql.conf file.

The parameter you want to change is:

statement_timeout = X

where X is in seconds;  set X = 0 to turn off the timeout completely (queries
will run as long as they need to).

        Andrew


8/29/03 2:35:04 AM, "Yudha Setiawan" <yudha@bonbon.net> wrote:

>
>
>  From:   "Yudha Setiawan" <yudha@bonbon.net>
>
>  To:     <pgsql-admin@postgresql.org>
>  Subject:[ADMIN] Can not deleted all record
>  Date:   Fri, 29 Aug 2003 13:35:04 +0700
>
>
>
>  Dear expert;
>
>  I've tried to update my record on stock's table.
>
>  but i failed "ERROR:  Query was cancelled."
>  was show up. But when i update on partial `till
>  all my record is changed, its work.
>
>  Anyway that's table have no rule,trigger and relation
>  with the other table. I've been thinking we must vacuum
>
>  this table.
>  But when i tried Vacuum with "Vacuum full verbose analyze
>  d_master.t_stock;" the "ERROR:  Query was cancelled." is
>
>  show up again.
>  Which part on my pg_hba.conf that indicates this error ????
>  or any idea about this ????
>
>
>  Please gimme your hand. I couldn't reset my table :(
>
>  GOD Bless You All and Bravo Postgres
>
>  ------------------------------------

Andrew Biagioni <andrew.biagioni@e-greek.net> writes:
> "ERROR:  Query was cancelled." usually means that the query is taking too long,
> and you exceeded the maximum time specified in your postgresql.conf file.

I doubt he has statement_timeout set, or it would have occurred to him
that he's shooting himself in the foot ;-)

My bet is that something external to Postgres is sending a SIGINT signal
to the backend.  There are a number of possible reasons for this, but
two mistakes I've seen newbies make are:
  * launch the postmaster from an interactive terminal, and forget to
    dissociate it completely from the interactive session (should use
    nohup and make sure that stdin, stdout, stderr are all redirected).
    Then any control-C typed in that session will zap the postmaster
    and all backends.
  * launch postmaster with non-infinite ulimit settings.  The backend
    trying to do a complete vacuum could be hitting limits on process
    runtime or amount of I/O.

            regards, tom lane