Title: RE: Shutdown Error (the old umount2 error)
Illegal seek sounds like a filesystem problem.
 
On the other hand, why aren't those processes you kill manually already down?  Also, fuser -mu /usr may provide useful info before you start killing processes.
 
HTH,
trever
-----Original Message-----
From: John H. Clark, III [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 19, 2002 7:11 AM
To: [EMAIL PROTECTED]
Subject: RE: Shutdown Error (the old umount2 error)

I'm having the same problem with a standard, unmodified RH 7.3. I verified that all the K and S links in runlevel 0 are present in response to an earlier suggestion from the list. And I don't have separate partitions in my /usr structure. Just a plain vanilla /usr structure.

However, the termination messages are the same, i.e.

Umount: /usr: Illegal seet
[Failed]
Init: No more processes....

Any ideas would be appreciated. This message is intended to keep the focus on this issue until some ideas start flowing.

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Benjamin R. Mohilef
Sent: Wednesday, September 18, 2002 11:54 AM
To: [EMAIL PROTECTED]
Subject: Shutdown Error (the old umount2 error)

Looking for suggestions on how to troubleshoot this thing more
easily:

Dual Athlon MP 2100+ workstation running RH Rawhide { a super
modified 7.3 } . IDE hard drive. The directories /usr and /usr/local
are on separate partitions for size reasons. The start for run level 0
is killall and reboot. With the help of a hapless passerby, I
compared level 6 Kxx symlinks to files in init.d and they are all
there with no redundant entries. Runlevel 0 has killall, halt.

The machine works great for numericals and simulations.  Problem
is when we shut it down. Almost acts like it's doing a cd to /usr 
before final shutdown. We get the familiar:

~~~~~~~~~~~~
Unmounting file systems:  umount2:   Device or resource busy
umount: /dev/hda3:  not mounted
umount: /usr: Illegal seek
    [FAILED]
INIT: no more processes left in this runlevel
~~~~~~~~~~~~


HOWEVER, it shuts down properly if I go into maintenance mode
and do:

~~~~~~~~~~
su                              { must be root }
/sbin/shutdown 0                { go to maintenance mode }
killall -9 minilogd             { one of the services affecting /usr }
killall -9 ntpd                 { another service affecting /usr }
killall -9 cupsd                        { last service affecting /usr }
umount /usr/local               { required to get rid of /usr }
umount /usr                     { unmount  /usr }
poweroff                        { or reboot -- works properly }
~~~~~~~~~~

For some reason I am missing something obvious. What have I
missed here?



Reply via email to