[1670] in SIPB_Linux_Development
Re: login takes me 13 minutes
daemon@ATHENA.MIT.EDU (Joanne M Mikkelson)
Tue May 13 17:50:26 1997
To: Eric Mumpower <nocturne@MIT.EDU>
Cc: linux-dev@MIT.EDU
From: jmmikkel@MIT.EDU (Joanne M Mikkelson)
Date: 13 May 1997 17:48:47 -0400
In-Reply-To: Eric Mumpower's message of Mon, 12 May 1997 11:39:48 -0400
> > Bleah. This is bad. It should work. Maybe this is a problem with
> > the PAM stuff? Do you have a password entry in /etc/passwd with an
> > encrypted password. (I'm sure you do, just checking. ;)
Yes, I do have an /etc/passwd with an encrypted password. :)
> I have had the same problem. I read through the login manpage, and found it
> wholly useless. My solution (read that "workaround") was to continue using
> login.krb, and simply to wait several minutes for a login.
Mine, too.
At the risk of sounding like a complete idiot, I have found out what
the problem is: named. The reason why I didn't expect it to be a
problem is that I *thought* I had named running without net before,
but I'm not sure. Thing is I didn't want to nuke all the network
configuration, for when I bring up ppp, and yes, I know I can fix up
all the scripts in /etc/sysconfig/network to set everything up from
ifup-ppp, or whatever, but I was assuming I could hobble along
without doing so, because I could before. I guess this isn't really
any different in the end from Eric's fix, in terms of effort. It's
not that I won't do it, I'm just out of the habit of wasting time on
my computer. :-)
Also, I wouldn't have really guessed that "Login timed out after 300
seconds" would be at all related to name service, but hey... Killing
named (or rather, not starting it at all) does reduce login time back
to a second or two.
> It's known that you will have problems with stock Linux-Athena
> if you're not on the net. We really should see if there's
> a way to work around this.
Could there maybe be a different rpm which deals properly with these
issues? It would be non-default, naturally, but it could be there
with its little description, and be mentioned in the installing
document. I don't know enough about rpm's to know if this is a pain.
> Yes, this is an annoying problem that we're aware of.
> I'm not sure what a good solution is. Maybe have
> the athena startup script exit if gettime times
> out (making gettime so it times out, of course... ;)
Well, I'm trying to find a reason why I shouldn't just go fix gettime
next time I'm punting my real work. The fact that it's been known to
do this for years makes me wonder if there isn't a reason for it. I
sure can't think of one myself. :)
Joanne
