autofs stops to work

autofs stops to work

Post by Johannes Ni » Wed, 14 Jul 1999 04:00:00



I have a running workstation wich is NFS client and also server (SMP
Pentium II / 2.2.5 Kernel/ SuSE 6.0 Distribution). Normally everything
works well.

Sometimes, about every few days, I am not able to mount local files
with nfs/automounter anymore. I get "no such file or directory".

This happens also with exported files, and also with files mounted via
NFS remotely from IRIX and Linux Hosts..

When I do

telinit 1

....

telinit 2

which shuts down autofs, nfs and nfsserver, this will help sometimes
but frequently not.

I get this warnings in the syslog:

Jul  9 15:04:39 scheat automount[28463]: shutting down, path = /data
Jul  9 15:04:39 scheat automount[28482]: shutting down, path = /-
Jul  9 15:04:42 scheat automount[28881]: starting automounter version 3.1.3, path = /vol, maptype = yp, mapname = auto.vol
Jul  9 15:04:42 scheat automount[28881]: Map argc = 2
Jul  9 15:04:42 scheat automount[28881]: Map argv[0] = auto.vol
Jul  9 15:04:42 scheat automount[28881]: Map argv[1] = hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28881]: parse(sun): init gathered options: hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28892]: starting automounter version 3.1.3, path = /data, maptype = yp, mapname = auto.data
Jul  9 15:04:42 scheat automount[28892]: Map argc = 2
Jul  9 15:04:42 scheat automount[28892]: Map argv[0] = auto.data
Jul  9 15:04:42 scheat automount[28892]: Map argv[1] = hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28892]: parse(sun): init gathered options: hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28903]: starting automounter version 3.1.3, path = /home, maptype = yp, mapname = auto.home
Jul  9 15:04:42 scheat automount[28903]: Map argc = 2
Jul  9 15:04:42 scheat automount[28903]: Map argv[0] = auto.home
Jul  9 15:04:42 scheat automount[28903]: Map argv[1] = hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28903]: parse(sun): init gathered options: hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28911]: starting automounter version 3.1.3, path = /-, maptype = yp, mapname = /etc/auto.direct
Jul  9 15:04:42 scheat automount[28911]: Map argc = 2
Jul  9 15:04:42 scheat automount[28911]: Map argv[0] = /etc/auto.direct
Jul  9 15:04:42 scheat automount[28911]: Map argv[1] = hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28911]: parse(sun): init gathered options: hard,intr,nosuid,bg
Jul  9 15:04:42 scheat automount[28881]: >> mount: automount(pid28881) already mounted or /vol busy

Jul  9 15:04:42 scheat automount[28881]: >> mount: automount(pid28881) already mounted or /vol busy
Jul  9 15:04:42 scheat automount[28881]: >> mount: according to mtab, automount(pid179) is already mounted on /vol
Jul  9 15:04:42 scheat automount[28881]: /vol: mount failed!

When I reboot, this problem disappears frequently but it may appear
again rapidly.

When I have remote filesystems mounted via fstab or manually, there
are no problems with them. The problems thus seems to be with autofs.

This happens also with an No-SMP Kernel.

The problem seems to be much more frequent if various mount points,
local or remote, refer to the same filesystem, e.g.

auto.home:
orion                   orion:/mnt/homes
jork                    orion:/mnt/homes/jork

Have to mention that there is a NIS auto.master entry which refers to
a direct mount for some SUN clients. Could it be that this drives
Linux autofs crazy ?

auto.master:
/etc/auto.direct        -hard,intr,nosuid,bg

I have been searching around now for two months and would appreciate
greatly any hints.

Exist there some resource for fixed bugs of the Linux Kernel or bugs
of software close to the Kernel ?

Johannes

--------------------------------------------------

http://www.uni-oldenburg.de/~jnix

Arbeitsgruppe Medizinische Physik
Universit?t Oldenburg
-------------------------------------------------------------
somebody knows a nice joke if Linux doesn't works perfectly ?
-------------------------------------------------------------