NFS crash with 2.1.129

Stephen Beynon (s8beyno@uk.ibm.com)
Thu, 26 Nov 1998 17:19:40 +0000


I have a repeatable problem where NFS causes the machine to lockup. I have
tested 2.1.127 and 2.1.129.

The NFS server is a VMS 6.2 machine (the VMS NFS server seems to be poor so
this is almost certainly a factor. )

I have the VMS system disk mounted, and when I cd to sys0.dir (I think that
sys0.dir on VMS contains some files with an unusual structure so I am sure that
this is not a coincidence) I receive the following error :

__nfs_fhget: inode 11 busy, i_count=2, i_nlink=2
nfs_free_dentries: found //sys0.dir, d_count=0, hashed=1
nfs_dentry_delete: //sys0.dir: ino=11, count=2, nlink=2

if I then try an ls or command line completion I receive the following errors
repeated at constantly at high speed. The machine is now unresponsive, Sysreq
will not even let me sync the disks, but it will let me change the console log
level.

NFS: inode 15 busy, no aliases
__nfs_fhget: inode 15 busy, i_count=1, i_nlink=1
__nfs_fhget: inode 15 still busy, i_count =1

I also have another nfs problem under 2.1.x that was not present under 2.0.x.
Under VMS directory files have a .dir extension. The VMS NFS server allows
these extensions to be omitted when changing directory so the following 2
commands should be equivilent

cd /mnt/users/stephen/temp
cd /mnt/users.dir/stephen.dir/temp.dir

unfortunatly if I try the first command I can perform an ls, but the getcwd
call fails ! I would guess that this is related to the dcache changes in the
2.1 series.

Any advice would be appreciated. Let me know if you need any more info
(tcpdump does not seem to work properly on the IBM token ring driver so packet
dumps may be a problem)

please CC me in any reply

Thanks

Stephen

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/