mountd not responding in 2.2.3

Godmar Back (gback@cs.utah.edu)
Tue, 9 Mar 1999 18:00:33 -0700 (MST)

Hi, I'm not really sure this is the right mailing list for this,
but the problem occurred after updating to 2.2.3 on my RH 5.2 box.
It's probably not a kernel problem, but it did show up for me in
conjunction with installing a new kernel.

So, if you feel this is the wrong mailing list for that, take replies
to "unfsd@monad.swb.de" only and please cc me since I don't know where
unfsd@monad.swb.de is archived.
Thank you for your attention.

I'm just posting this in case it rings a bell for anyone.

Having tackled the NFS client side (see separate thread), I'm now looking
at the server side. I'm using the userland "Universal NFS Server 2.2beta40"

After I reboot, RH 5.2 starts rpc.mountd and rpc.nfsd just fine, but
mountd does not respond.

rpcinfo shows that portmap registered it just fine:

gback@marker [7](~) > rpcinfo -p peerless
program vers proto port
100000 2 tcp 111 portmapper
100000 2 udp 111 portmapper
100007 2 udp 803 ypbind
100007 2 tcp 805 ypbind
100005 1 udp 635 mountd
100005 2 udp 635 mountd
100005 1 tcp 635 mountd
100005 2 tcp 635 mountd
100003 2 udp 2049 nfs
100003 2 tcp 2049 nfs
300019 1 tcp 915 amd
300019 1 udp 916 amd
100021 1 udp 1024 nlockmgr
100021 3 udp 1024 nlockmgr
100021 1 tcp 1024 nlockmgr
100021 3 tcp 1024 nlockmgr

Portmap works just fine too, I can talk to other services:

gback@marker [5](~) > rpcinfo -t peerless 100007
program 100007 version 2 ready and waiting

I cannot talk to mountd, however:

rpcinfo -t peerless 100005
rpcinfo -t peerless 100005 1
rpcinfo -t peerless 100005 2
rpcinfo -u peerless 100005
rpcinfo -u peerless 100005 1
rpcinfo -u peerless 100005 2

all fail with

rpcinfo: RPC: Timed out

Now when I type "netstat -t inet" on the machine on which rpc.mountd
runs, I see this:

Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 45 0 peerless.cs.utah.ed:635 marker.cs.utah.edu:1249 CLOSE_WAIT

This looks to me as though rpc.mountd does not pick up the data.

Note that this happened already with the 2.2beta37 version that came
with RH 5.2; upgrading to beta40 didn't make a difference.
[root@peerless init.d]# rpc.mountd -v
Universal NFS Server 2.2beta40

Now comes the funny part.
I stop and restart nfs (both rpc.nfsd and rpc.mountd) like so:
cd /etc/rc.d/init.d
./nfs stop
./nfs start

Nothing changes (i.e., rpcinfo still fails.)

Now, I manually look up the pid of rpc.mountd:

[root@peerless init.d]# pidof rpc.mountd
565

Kill it:

[root@peerless init.d]# kill 565

and start it manually like so:

[root@peerless init.d]# rpc.mountd

and voila, mountd responds.

gback@marker [15](~) > rpcinfo -t peerless 100005
program 100005 version 1 ready and waiting
program 100005 version 2 ready and waiting

and subsequently, it exports the filesystems like it should.

So, my question is this:

What weird interaction of kernel version, NFS server version and RedHat
setup do I fall victim here?

A second question: would people recommend I use the in-kernel nfsd instead?
Does this in-kernel nfsd also use an in-kernel mountd? What's the future
outlook here?

Thanks,

- Godmar

-
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/