i use the knfsd-981014 package under LinuxPPC 2.1.119.
I patched the kernel sources with :
knfsd-981014/linux-2.1.1xx.diff
knfsd-981014/root-2.1.1xx.diff
knfsd-981014/sunrpc-2.1.123-1.patch
I recompiled the kernel and it seem to be ok.
I have on two (mtx0,mtx1) machines the /etc/exports :
/home (rw)
I installed the RedHat style init script nfs and after the boot
sequence on each machine i have these daemons running (i asked for 4 nfsd) :
bin 132 0.0 0.0 0 0 ? SW 11:13 0:00 portmap
root 169 0.0 0.0 0 0 ? SW 18:22 0:04 rpc.kstatd
root 179 0.0 0.0 0 0 ? SW 18:22 0:00 rpc.kmountd --no-nfs-version 3
root 188 0.0 0.0 0 0 ? SW 18:23 0:00 (nfsd)
root 189 0.0 0.0 0 0 ? SW 18:23 0:00 (nfsd)
root 190 0.0 0.0 0 0 ? SW 18:23 0:00 (nfsd)
root 191 0.0 0.0 0 0 ? SW 18:23 0:00 (nfsd)
root 194 0.0 0.0 0 0 ? SW 18:23 0:00 (lockd)
root 195 0.0 0.0 0 0 ? SW 18:23 0:00 (rpciod)
I mount mtx0:/home on mtx1 at /home, so i log into mtx0 and mtx1 with
my home directory at /home :
mtx1 root # mount -t nfs mtx0:/home /home
Here i have a program that uses locks to access a shared file
over NFS. I tried the testlk tool and locks seem to work for
me. The program i use is also writing some data to a result file.
But, after a while i have my process on the mtx1 (NFS client) that are blocked,
when i do a ps here is what i have :
USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND
delchini 651 0.0 1.4 1856 900 p0 D 12:53 0:00 synctest
As you see, it is in the "D" status, if i try to kill it, it doesn't
die. I must reboot to let it disapear.
I think this process is blocked on an IO that never ends, no ?
May be i should use some flags on the server and/or the client
to avoid these strange things ?
Any help will be apreciated.
Ugo (delchini@lpnp09.in2p3.fr).
-
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/