Re: 7.3 LOCK TABLE problem

Neil, Tom, and anyone else following the thread: Thank you for your time and
suggestions. I feel like I never say that enough when participating in a
development community. Andrew Sullivan was saying earlier how much he enjoyed
participating in the PostgreSQL community, and using such a well-maintained and
supported software product. I agree with him completely.

To my frustration, I was not able to reproduce this problem on a clean install
of PostgreSQL 7.3.1. I can reproduce it only on the production machine. In
other words, it is MY problem, not PostgreSQL's. It seems as though the problem
is localized to my particular installation, which originated as a 7.0 server
and has been upgraded to 7.1, and 7.2 (which shouldn't matter). I checked the
PgSQL.pm versions on both the production machine and the "clean install"
machine. The versions match. I'm at a complete loss to explain why this is
happening! I am also at a complete loss to provide you with a bug that can be
reproduced (save boxing up and sending you our production server).

I read the new 7.3 admin guide. I'm embarrassed to say that I'm not sure what I
might be looking at if I enabled these options... Which ones would not help me
to diagnose the subsystem that's causing my problem?

server_min_messages = debug5 ?
client_min_messages = debug5 ?
debug_assertions = true ? (what ARE assertions?)

(after compiling with LOCK_DEBUG)
trace_locks = true ?
trace_userlocks = true ?
trace_lwlocks = true ?
debug_deadlocks = true ?
trace_lock_oidmin = 16384 ?
trace_lock_table = 0 ?

I'm fine with spending days taking apart the whole software setup. But who has
days to spend on a problem like this? :) I'm hoping for a clue to get me
digging in the right desert.

CG


--- Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Chris Gamache <cgg007@yahoo.com> writes:
> > If I can reproduce this on a clean install of PostgreSQL, then what next?
>
> Send a complete reproducible example (not an off-the-cuff script you
> haven't actually tested ;-)).
>
> Like Neil, I know of no server-side reason for the behavior to have
> changed since 7.2.  I'm wondering if there's something different in the
> client-side libraries.

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com