Return-Path: mao
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA10687; Thu, 13 Aug 92 13:55:26 -0700
Message-Id: <9208132055.AA10687@postgres.Berkeley.EDU>
From: <mao@postgres.berkeley.edu>
Subject: Re: size for btree indices
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu
In-Reply-To: Your message of Thu, 13 Aug 92 13:48:43 PST.
             <9208132048.AA10626@postgres.Berkeley.EDU> 
Date: Thu, 13 Aug 92 14:08:04 PDT

In message <9208132048.AA10626@postgres.Berkeley.EDU> you write:

>     My conclusion, therefore, is that the problem is in the size of the
> relation and not in the data itself.  Is there a limit on the size of
> a relation which can be indexed?  Is this what you intend?

no such limit exists.  the postgres backend is crashing when it tries
to build your index.  there were some bugs reported in the text_ops
operator class in 4.0, which have been fixed in the upcoming 4.0.1
release.  it's hard to say whether these would have caused the crash
you report, but it's possible.  if you could send us a stack backtrace
(the core file is in $POSTGRESHOME/data/base/your_db_name) we might
be able to say something more definitive.

					mike olson
					project sequoia 2000
					uc berkeley
					mao@cs.berkeley.edu