[174] in testers
Re: new attach
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Mon Jul 24 22:42:42 1989
Date: Mon, 24 Jul 89 22:41:47 EDT
From: Theodore Ts'o <tytso@ATHENA.MIT.EDU>
To: John T Kohl <jtkohl@ATHENA.MIT.EDU>
Cc: testers@ATHENA.MIT.EDU
In-Reply-To: John T Kohl's message of Mon, 24 Jul 89 20:40:38 EDT,
Reply-To: tytso@ATHENA.MIT.EDU
Date: Mon, 24 Jul 89 20:40:38 EDT
From: John T Kohl <jtkohl@ATHENA.MIT.EDU>
X-Us-Snail: MIT Room E40-351, 1 Amherst St., Cambridge, MA 02139 USA
The new attach does not use the FQDN in its RPC requests to the mount
daemons, which leads to "permission denied" messages from servers which
check the IP address against the hostname.
Yes, but I think that will force people to put the FQDN in /etc/exports,
possibly causing lossage since people won't know that. I thought about
doing such a thing, since I've installed attach on a LCS host, but
thought it would cause some nasty unexpected lossage for Athena users.
Besides, that's what "spoofhost" is for! :-)
On a more serious note, /etc/mount does not put the FQDN in rpc requests
(which means that AI and LCS people have a difficult time NFS mounting
Athena filesystems). Do we really want to break comformance with this
``feature''? It will mean that people will have to put both forms of
the hostname in /etc/exports.
- Ted
