General Bug Report: psql does report failed SQL commands as executed - Mailing list pgsql-bugs

From Unprivileged user
Subject General Bug Report: psql does report failed SQL commands as executed
Date
Msg-id 199907190847.EAA00964@hub.org
Whole thread Raw
List pgsql-bugs
============================================================================
                        POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name               : Harald Laabs
Your email address      : laabs@uni-duesseldorf.de

Category                : unknown
Severity                : critical

Summary: psql does report failed SQL commands as executed

System Configuration
--------------------
  Operating System   : solaris 2.6

  PostgreSQL version : 6.4.2 and 6.5

  Compiler used      : gcc 2.8.1

Hardware:
---------
solaris on sun sparc
also tested: Linux 2.2.5 on Pentium, 32M

Versions of other tools:
------------------------


--------------------------------------------------------------------------

Problem Description:
--------------------
When inserting many (>1000) SQL INSERT commands in psql in a short time some of them are not executed while they are in
thelog-file. 
Because of an error message "query buffer max length of 20000 exceeded" I put a "sleep 60" between blocks of 200
INSERT's.
Now there are about 100 commands not executed every 2000 Lines of SQL commands.
I experienced the same problem on a Linux 2.2.5 machine with postgres 6.4.2 and 6.5 at home.

--------------------------------------------------------------------------

Test Case:
----------
Create a new database (numismatik)
get http://www-public.rz.uni-duesseldorf.de/~laabs/*
psql numismatik <definition.sql
split -l 200 pool00.sql
pipe the (first) parts in psql (with about a minute between the parts, or "query buffer max length of 20000
exceeded"???)-> use insert_script &>insert.log & 
there should be one entry in mr_dat for every number between 1002 and 2000 (or lower if you insert only the first
parts,but every number between the lowes and highest MUST appear, but 1088-1093 are missing here)  

--------------------------------------------------------------------------

Solution:
---------


--------------------------------------------------------------------------


pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: [BUGS] 'Default' troubles again. This time with time :)))
Next
From: Unprivileged user
Date:
Subject: General Bug Report: ODBC translation of Table/Field Names