Messages in this thread | | | Subject | Re: the way 2.1 shuts down compared to 2.0 | Date | Wed, 20 May 1998 11:49:58 +0200 | From | Olaf Titz <> |
| |
In article <9805190403.AA13246@gnu.sdsp.mc.xerox.com> you write: > When I shutdown 2.0.33 or 2.1.*, I get: > Unmounting file systems > RPC portmapper failure > RPC unable to receive > umount:/usr/local device busy > umount:/usr/gnu device busy > umount:/usr/X11R6 device busy > While I have amd running off /usr/local, I don't see anything off > /usr/gnu. > What is annoying, is 2.0 just reboots anyway, but > 2.1 sits in a loop and just prints: > RPC sendmsg returned 101
But you _want_ to unmount these file systems cleanly. A "busy" in the shutdown script means something's seriously wrong in your setup.
I had this problem recently, with my big /var disk not getting unmounted and fsck'd on every reboot. I hadn't any idea why this happened, so I inserted an "lsof > /debug" jut before the umount in the shutdown script and tried to analyze its output after rebooting.
It turned out that the Sun JDK, which I had installed on the /var disk, came with its own copy of libc.so.5.4.something deep inside an obscurely nested path and ldconfig happily pointed every program out there, including /bin/sh, to use that instead of the copy in /lib. This also happened with another library (I think it was libdl). Blech. (Fix, of course, was to remove all those redundant libraries.)
olaf
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |