I'm having a very nasty problem with the kernel space NFS server.
I have four machines inter-connected by NIS and NFS (mounted by
amd). One of them is connected by a slow (36.6 kbps) ppp link.
One of the other three provides smb services with files in the local
filesystem and in the other servers (for the user's homes).
Everything works fine except when the ppp link goes down with the
nfs filesystem mounted in the smb server. The processes using a
file in the remote server hung normally waiting for IO, and if you
attempt to access any resource of that server by an "ls" or a "cat"
the automounter responds "resource temporarily unavailable" and
returns to the shell. But if you do a "df" the process hangs and
don't responds to CTRL-C. (note: the nfs is mounted hard,intr)
Something similar happens with any new smbd process, with every
share acceded!!
I connect with smbclient to a local share and do an ls, it shows me
the list, the free space left in the share and then the smbd hungs
waiting for IO. So I think it must be stating all the filesystems and
getting the same behaviour as "df".
Is this a bug in the knfsd or am I doing something wrong?
My configuration is:
smb server:
kernel 2.2.5-15
samba 2.0.4b-19990527
am-utils 6.0.1s11-1.6.0
knfsd 1.2.2-4
ypbind 3.3-20
Remote server:
kernel 2.2.5-22
autofs 3.1.3-2
samba 2.0.3-8
knfsd 1.2.2-4
ypbind 3.3-20
--- * Martin Ferrari- 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/