Return-Path: pg_adm@postgres.berkeley.edu
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA11607; Mon, 7 Dec 92 04:13:57 -0800
Date: Mon, 7 Dec 92 04:13:57 -0800
Message-Id: <9212071213.AA11607@postgres.Berkeley.EDU>
From: Leonard Norrgard <vinsci@brando.uwasa.fi>
Subject: initdb problems
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu

I'm building Postgres 4.0.1 on a sparc with OS/MP 4.1A.3 Export from Solbourne
(functionally equivalent to Sun's, except perhaps less bugs, I'm told).

Everything goes well until initdb time.  Initdb dies as follows:

> initdb: verbose mode on
> initdb: looking for postgres...
> initdb: found /usr/postgres/bin/postgres
> initdb: looking for template files...
> initdb: using /usr/postgres/files/local1_template1.bki
> initdb: using /usr/postgres/files/global1.bki
> initdb: creating SHARED relations in /usr/postgres/data
> initdb: creating template database in /usr/postgres/data/base/template1
> WARN:Dec  7 13:41:19:cannot open          pg_proc
> WARN:Dec  7 13:41:19:must OPEN RELATION before INSERT
> 
> WARN:Dec  7 13:41:19:must OPEN RELATION before INSERT

Lots of errors follow, but the key seems to be in "cannot open
pg_proc".  What might be the cause of this?  I've tried doing initdb
both as user postgres and root, but the result is the same.

Postgres runs with uid 7 on this machine.  The "postgres installation
instructions" document says that the "relase notes" document has info
on how to get around this, but it has not.  The older versions of the
release notes mention ways to do this, but no longer correspond to the
actual source code involved.  Besides, looking at the current source
code, it seems that it handles uid:s other than 6 automatically these
days.  Time to update the installation instructions?

I tried reporting this to bug-postgres@postgres.berkeley.edu a couple
of weeks ago and have so far heard nothing (except for the automated
reply).  What is the intended use for bug-postgres, really?  Shouldn't
reports to that address _at_least_ end up in pub/postgres-v4r0r1.bugs
on postgres.berkeley.edu? That file still has an August datestamp.

-- Leonard
