On Mon, 26 Jan 1998, Colin Wood wrote:

> SamMaEl wrote:
> > 
> > 	Hey folks, I have an unusual problem. Well, things have been
> > chugging along normally for me, but tonight I was planning on finally
> > upgrading from 1.3_BETA to 1.3. I already had the files downloaded and I
> > was all set to go. Well, tonight I rebooted the system after it had been
> > up for about 2 weeks, and things were just fine. Then... I downloaded MI/X
> > and was playing around with that on one of my MacOS machines, with the
> > NetBSD machine supplying the xterms and the window manager. THAT went
> > fine. 
> 
> good.
>  
> > 	Well, I wanted to be able to use my .twmrc file for when my other
> > users ran xterms and stuff on the MacOS machine, so I copied my
> > /root/.xinitrc file to /home/sammael the home directory for my user
> > account. I also copied it to /etc/skel so when I created new users, they
> > would automatically have the files for use. The ONLY other thing I did was
> > I copied my binary sets over to my root partition. (they were sitting on
> > another partition, to save disk space on my root)
> 
> you didn't extract any of these files?  how much room do you have left on
> your root partition?  it wouldn't happen to be full would it?

	No... I had not extracted any of the files yet... and I had only
copied base.tgz over when it started happening. 

> > 	Then, I started getting a message on the console from sendmail
> > complaining about a bad aliases database. I tried to run newaliases, but
> > it complained that it didn't have permission to open the sendmail.cf file.
> > And then, my normal users could not login properly, because their csh
> > would not exec their .cshrc files properly on login. Also, the
> > uid/loginname mapping was screwed up too. (files showed up as
> > sammael:sammael when root did an ls -l, but when sammael did it, I got
> > 1000:1000) I also could not enter my home directory at all.... sammael was
> > stuck in / and any other file systems that were readable by anyone.
> 
> my only guess on this one right now is that you have somehow managed to
> competely fill up your root partition when you copied the new distribution
> sets onto it, and now a number of file accesses have truncated or
> completely blow away several rather important files in /etc.  Check the
> contents of things like /etc/master.passwd and make sure the spwd.db? file
> isn't truncated as well.

	That would make sense... but my file system is NOT full, accoring
to 'df'. It says that /dev/sd0a (root) is 74% full, with over 100 Megs
free. And /dev/sd0g (/hobbes) has over 100 MB free as well.

> 
> > 	Has anyone else seen something like this happen? Right now I'm
> > doing the 1.3 upgrade, hoping this will fix it. I'm in the habit of using
> > su now, so having to log in as root all the time would/will be a pain. Not
> > to mention it could screw up my file permission scheme. (I have alot of
> > different files I have owned by different users, even though I'm the only
> > one that uses the system, generally... just a habit ;-))
> 
> a 1.3 upgrade might fix most of the problems, but if your root partition
> is really full then untarring over everything may leave you with a bunch
> of truncated files....quite messy!

	No.... master.passwd and passwd look fine. Though, I did not wipe
them on my upgrade... I copied the 'originals' over the ones that etc.tgz
copied over them.

	Now, I just got another error I've never seen, this time from
identd.	It gave this error:

identd[511]: /etc/pwd.db" Permission denied
identd[511]: getpwuid() could not map uid (0) to name

	I logged into a chat program that I run, which runs an ident
client I wrote. And, it does an ident lookup... which normally does not
give this error ;-) 

	Another bizarre error I got, was when I rebooted the system after
the upgrade....

Jan 26 09:36:15 hobbes root: WARNING: $xfs is not set properly.
Jan 26 09:36:18 hobbes root: WARNING: $xdm is not set properly.

	I'm not RUNNING XDM, or XFS, and I actually didn't upgrade my /etc
until after I upgraded everything and rebooted, so I don't know what
caused THOSE errors...

	*sigh* all this and I've been up for like.... too long ;-)

	Ryan

	P.S. Does anyone know what changes have been made to grf_iv.c?
Michael Zucca sent me a patch to hopefully get my q630 machine's internal
video going properly, but when I patched it (with patch) I got some
errors... should I just go ahead and download the newest one, or will that
cause more problems? The sources I have are... hmmm... the newest file in
/sys/arch/mac68k/dev is from Dec 1st, and the grf_iv.c gives a date of Nov
10. Should I just download the newest one? Michael?

-----
HELO... my name is rewt... you have SIGKILLed my father... prepare to vi!