I would appreciate any pointers to how to debug this:

"Suddenly" NFS doesn't seem to be working on one of my machines. I have a fairly extensive home network with 8 networked computers. Several of them have shares which are automounted via NFS.

The machine in question is my "cutting-edge" box running Debian-unstable, and is always kept updated. Recently, I have not been able to access my NFS mounts from this box.

The network is up, pinging works well, both from and to the box. The mounts are appropriately mounted as per the "mount" command. However, I cannot cd to the mounted FS.

I have:

>> done apt-get dist-upgrade which went without any hassle

>> removed and reinstalled nfs-kernel-server and nfs-common

>> powered down the box several times

>> changed all the connections on the router and switch around, and power- cycled those several times.

>> checked ifconfig (in desperation)


I just don't know where to look further - does any one have any suggestions?

Thanks for any replies
Steve

What do your logs tell you on both the client and NFS server your mounting from? That's always usually the first step in finding out what exactly is wrong. And don't always trust the mount command in such situations, it can be deceiving.

Also, have you tried not using autofs to mount these shares and manually mounted the shares?

Ah .... the LOGS, who would have thought of that?

/var/log/messages shows this

( I have not shown all the entries, there are more of them, the ***s are edited in )

Jan 23 12:08:48 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:09:11 *** kernel: eth0: link down
Jan 23 12:09:13 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:09:14 *** kernel: eth0: link down
Jan 23 12:09:16 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:09:33 *** kernel: eth0: link down
Jan 23 12:09:34 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:09:37 *** kernel: eth0: link down
Jan 23 12:09:38 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:09:39 *** kernel: eth0: link down
Jan 23 12:09:40 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:10:05 *** kernel: eth0: link down
Jan 23 12:10:07 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
Jan 23 12:10:11 *** kernel: eth0: link down
Jan 23 12:10:13 *** kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1

dmesg, of course, shows the same.
So this means that I have an unhappy network card, right?
Strange thing is, my internet connection is working just fine, but I guess I do not need a "persistent" connection for the "stateless" internet, right?

/var/log/daemon.log
/var/log/user.log
/var/log/news/news.crit
/var/log/news/news.err
/var/log/news/news.notice
/var/log/debug

did not show much that seemed helpful.

So, I suppose the next step would be to swap out the network card, right? Do you know of any software that could help diagnose the NIC? The rest of the network seems to be doing OK, so I would think that any hardware issue would be on this machine and not on the router / switch.

Yes, I forgot to mention that. Manual mounting simply freezes the console.

Thanks for the help, trickykid
Steve

Hmmmm ... I'm also seeing similar entries in var/log/messages on the server, so the problem must be the router, then.

I'm going to simplify the network to only 2 computers, and see what happens.

Steve

OK, that did not fix things either.

This is the message that I am seeing in /var/log/messages on the client (192.168.0.10):

Jan 23 12:49:35 *** kernel: Inbound IN=eth0 OUT= MAC=00:40:ca:54:bf:7c:00:50:ba:50:95:eb:08:00 SRC=192.168.0.12 DST=LEN=44 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=TCP SPT=111 DPT=901 WINDOW=5840 RES=0x00 ACK SYN URGP=0

This is repeated many times

/var/log/messages on the host (192.168.0.12) is happier now.

00:40:ca:54:bf:7c: is the MAC for 192.168.0.10
00:50:ba:50:95:eb: is the MAC for 192.168.0.12

I am not sure what the 08:00 represents.

[edit]
When I add the rest of the network back, I see similar strings with the MAC for the other machines substituted in the above string
[/edit]


I tried 3 different switches ( 1 router, and 2 switches, all with only the 2 machines in question connected. I can still ping the host from the client and vice versa, and mounting from the client still hangs, although it does not hang the console. ( I rebooted both host and client again)

Any thoughts?

Steve