Hi folks - I'm seeing a really unusual problem when I try to access a nonexistent mount under /home - cd'ing into "/home/asdf" causes the automounter to get stuck in what appears to be an infinite loop (in the example below, something tried to access /home/lock and caused the problem):

Oct 15 14:13:55 seneca automount[16767]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16767]: failed to mount /home/lock
Oct 15 14:13:55 seneca automount[16769]: >> mount: brain-nfs:/ldisk/md0/home/lock 
failed, reason given by server: No such file or directory
Oct 15 14:13:55 seneca automount[16769]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16769]: failed to mount /home/lock
Oct 15 14:13:55 seneca automount[16771]: >> mount: brain-nfs:/ldisk/md0/home/lock 
failed, reason given by server: No such file or directory
Oct 15 14:13:55 seneca automount[16771]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16771]: failed to mount /home/lock
Oct 15 14:13:55 seneca automount[16773]: >> mount: brain-nfs:/ldisk/md0/home/lock 
failed, reason given by server: No such file or directory
Oct 15 14:13:55 seneca automount[16773]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16773]: failed to mount /home/lock
Oct 15 14:13:55 seneca automount[16775]: >> mount: brain-nfs:/ldisk/md0/home/lock 
failed, reason given by server: No such file or directory
Oct 15 14:13:55 seneca automount[16775]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16775]: failed to mount /home/lock
Oct 15 14:13:55 seneca automount[16777]: >> mount: brain-nfs:/ldisk/md0/home/lock 
failed, reason given by server: No such file or directory
Oct 15 14:13:55 seneca automount[16777]: mount(nfs): nfs: mount failure 
brain-nfs:/ldisk/md0/home/lock on /home/lock
Oct 15 14:13:55 seneca automount[16777]: failed to mount /home/lock

<...and on and on until /var/log/syslog fills up the disk...>

I'm not sure if this is the same bug mentioned in autofs-4.1.3-strict.patch. ???

Problem surfaced when I upgraded from 2.4.25 to 2.4.27, but it doesn't look like there were any autofs4 changes made between 2.4.25 and 2.4.27. If it makes any difference, I am using the SNARE C2 auditing patches for the kernel, but I don't think those would have any impact on autofs.

I'm seeing this problem regardless of what autofs package I use (SuSE 9.0 autofs3, autofs4.0.0-pre10, or the latest autofs-4.1.3 + patches from kernel.org).

I've upgraded util-linux to 2.12b and nfs-utils to 1.06, no luck. Have also upgraded nfs-utils server-side, no luck.

I also applied the autofs4-2.4.22-20040508.patch to 2.4.27 and haven't seen any change in behavior (the autofs-4.1.3-strict.patch mentioned a the kernel patch might be required if the problem didn't go away).

More details about my system:

automount:  Linux automount version 4.1.3
mount:  mount-2.12b

uname -a: Linux seneca 2.4.27 #4 Tue Sep 14 14:42:16 CDT 2004 i686 i686 i386 GNU/Linux
  (built again today with the autofs4 kernel patch, no change)

# /etc/auto.master
----
/home   /etc/auto.home
/proj   /etc/auto.proj
/data   /etc/auto.data
/info   /etc/auto.info
/scratch /etc/auto.scratch
----

# /etc/auto.home
----
* -fstype=nfs,rw,soft,intr brain-nfs:/ldisk/md0/home/&
----

Have also tried /etc/auto.home without ANY options, no difference.

Any possible solutions you all could throw out there would be greatly appreciated - I'm stumped!

Robert Giles
ARL:UT

_______________________________________________
autofs mailing list
[EMAIL PROTECTED]
http://linux.kernel.org/mailman/listinfo/autofs

Reply via email to