[2832] in Release_Engineering
Re: Problem with setpag and login/xlogin
daemon@ATHENA.MIT.EDU (Richard Basch)
Mon May 18 10:38:25 1992
Date: Mon, 18 May 92 10:37:44 -0400
To: Ezra Peisach <epeisach@MIT.EDU>
Cc: "Richard Basch" <basch@MIT.EDU>, epeisach@MIT.EDU, builder@MIT.EDU,
In-Reply-To: Ezra Peisach's message of Mon, 18 May 92 09:09:15 EDT,
From: "Richard Basch" <basch@MIT.EDU>
Date: Mon, 18 May 92 09:09:15 EDT
From: Ezra Peisach <epeisach@Athena.MIT.EDU>
Without looking at the code, I thought of another problem which this will
not satisfy. The scenario of logging into and then say telnetting back in.
You will have your homedir detached. I think what is really needed is a
mechanism that will try to authenticate (if there is reason to believe
the homedir is already mounted) and if that fails, attach.
The reason for this is that we don;t want the homedir to be detached if you
login twice. Inly the one that really did the attach should unmount.
Ezra
Yes, well, the "detach" mechanism is bad; I would advocate the change of
"detach $USER" to "fsid -u [-a|-f $USER]" upon logout; let the automatic
workstation cleanup deal with the rest (reactivate/crontab.local/etc).
Unfortunately, that is not a new problem -- it has existed forever, and
really only affects those of us with AFS homedirs (yes, I know...) and
those who have multiple logins. Of course, it is a problem that should
be solved given the migration coming up...
I'll be doing some more testing and come up with another mechanism that
works with Ultrix... I thought the device was 0,1, but Ezra proved me
wrong (it was random...)
-R
