Return-Path: postarch
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA12937; Fri, 28 Feb 92 11:34:12 -0800
Message-Id: <9202281934.AA12937@postgres.Berkeley.EDU>
From: postarch (Postgres Mailing Archive)
Subject: Re: Error: No response from the backend, exiting...
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu
Reply-To: mer@postgres.berkeley.edu
In-Reply-To: Your message of "Tue, 25 Feb 92 12:59:42 PST."
             <9202252059.AA26714@postgres.Berkeley.EDU> 
Date: Fri, 28 Feb 92 11:33:56 PST


you write:

> Last week I posted a question regarding this error.
> I guess I could not make myself clear to get some help,
> so let me try again. I'll really appreciate (and need)
> any clue you can come up with.
> During this time, i've realized that i was trying to
> run my program from different machines, and that was
> precluding postgres from using effectively its shared
> memory mechanism. so i tought that was the origin of all
> my problems. however, when i shifted to run (only) 2

As long as you are using the same host machine (i.e. the
same postmaster) this won't make any difference.

> copies of it in the same machine, the problem persists.
> this time the results are somewhat different. sometimes
> it works perfectly, allowing more than 1 program to
> access the db. other times, one program works properly
> and the other exits with the message:

This sounds like bugs in postgres rather than bugs with your
program.  It sounds like one of the backends is choking
and dying (hopefully dumping core).  Look in the directory
$POSTGRESHOME/data/base/<dbname> for a core fileaction before proceeding to use the involved classes.

Your assumptions are correct.  Postgres does two-phase locking, so if 
one program holds write locks the other must wait until the transaction 
commits or aborts before it can access the locked data.


Jeff Meredith
mer@postgres.berkeley.edu