[511] in arla-drinkers
Re: Pb getwd on afs linux
daemon@ATHENA.MIT.EDU (Gerald Britton)
Tue Jan 19 09:47:01 1999
From owner-arla-drinkers@stacken.kth.se Tue Jan 19 14:47:00 1999
Return-Path: <owner-arla-drinkers@stacken.kth.se>
Delivered-To: arla-drinkers-mtg@bloom-picayune.mit.edu
Received: (qmail 5357 invoked from network); 19 Jan 1999 14:46:59 -0000
Received: from unknown (HELO sundance.stacken.kth.se) (130.237.234.41)
by bloom-picayune.mit.edu with SMTP; 19 Jan 1999 14:46:59 -0000
Received: (from majordom@localhost)
by sundance.stacken.kth.se (8.8.8/8.8.8) id PAA17426
for arla-drinkers-list; Tue, 19 Jan 1999 15:39:17 +0100 (MET)
Received: from light-brigade.mit.edu (exim@LIGHT-BRIGADE.MIT.EDU [18.244.1.25])
by sundance.stacken.kth.se (8.8.8/8.8.8) with ESMTP id PAA17416
for <arla-drinkers@stacken.kth.se>; Tue, 19 Jan 1999 15:39:12 +0100 (MET)
Received: from gbritton by light-brigade.mit.edu with local (Exim 2.10 #1)
id 102cJ6-0001xK-00; Tue, 19 Jan 1999 09:39:04 -0500
Message-ID: <19990119093903.A7220@light-brigade.mit.edu>
Date: Tue, 19 Jan 1999 09:39:03 -0500
From: Gerald Britton <gbritton@mit.edu>
To: "Brandon S. Allbery KF8NH" <allbery@kf8nh.apk.net>
Cc: arla-drinkers@stacken.kth.se
Subject: Re: Pb getwd on afs linux
References: <19990119081901.A2389@light-brigade.mit.edu> <199901191434.JAA27325@rushlight.kf8nh.apk.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.93.2
In-Reply-To: <199901191434.JAA27325@rushlight.kf8nh.apk.net>; from Brandon S. Allbery KF8NH on Tue, Jan 19, 1999 at 09:34:39AM -0500
Sender: owner-arla-drinkers@stacken.kth.se
Precedence: bulk
On Tue, Jan 19, 1999 at 09:34:39AM -0500, Brandon S. Allbery KF8NH wrote:
> Because that works by inode number. The problem being that AFS uses a set
> of four 32-bit values, which have to be hashed into a single 32-bit value;
> the result is *not* guaranteed to be unique, so ugly things can happen with
> programs that rely on unique inode numbers. Such as pwd, and such as
> Solaris's ld.so which goes bonkers if two shared libraries on AFS hash to
> the same inode number....
If this is the case, how does Transarc AFS deal with it? You never run into
this problem using their AFS client.
-- Gerald
