Thanks Gavin for your suggestions. It took me a while to work through your list. -I had been using a cat5e cable. Swapped it out with others, but same result. -/sbin/ifconfig was normal. /var/log/syslog had nothing much. /var/log/messages had one line that indicated a successful request for NFS services, but nothing else. -pinging showed no problems -Installed Ethereal (wireshark?) on server and sniffed during the thinclient boot. DHCP fine and a bunch (100+) of NFS packets back and forth. Not equipped to understand them, but nothing Ethereal marked as errors. -I installed a new 6.10 Ubuntu workstation, connected @ gigabit to Edubuntu 6.10 server (it's the DHCP server) and set up a Samba share (I'm more familiar with samba.) I could successfully tranfer a large file. -I setup NFS share on server and when I mounted it on the workstation it never came back to the command prompt (using the gigabit switch.) Rebooted all and tried with 100Mbps switch. Again, it took over 60 seconds for the client command prompt to return, but the NFS did work and I was able to transfer the same 2MB file in a few seconds. You may recall that I could swap switches during the thinclient boot (from gigabit to 100) and it would continue, so I tried the same thing using this workstation. Rebooted all and mounted NFS share while using the 100Mbps switch. Then, switched to the gigabit switch and tried the file transfer. It worked just as well! (FWIW, I sniffed both transfers. There were some TCP windowing errors in both that I would consider normal, but only the gigabit transfer had packet out-of-order TCP errors.) So, my tenative conclusion is the client/server cannot establish a NFS share while the gigabit switch is in place, but it can employ one when it has already been established. Not much good when you have to boot a thinclient.  Does this information help? Any suggestions? Thanks, Scott <div>On 1/5/07, Gavin McCullagh <<a href="mailto:gmccullagh@gmail.com">gmccullagh@gmail.com</a>> wrote:<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> The behaviour suggests your client cannot get to the nfs server while you're on the 1Gb port but can when you're on the 100Mb port.  That's quite strange.  It's certainly not fundamental to gigabit use -- we have a server on 1Gb and it works fine. It sort of looks like there's something wrong with the gigabit link but it's strange that you say it works in other situations.  Perhaps you're getting packet loss but this is only obvious in certain applications whose error checking is not so good (ie UDP ones such as NFS). A few guesses would include: - a faulty or below spec cable (for GBit you should need Cat5E/Cat6) - software errors at gigabit speeds - a slightly bad gigabit connection on the switch - some ACL rule blocking nfs traffic on the gigabit port of that switch    (if the switch is even capable of that) You could try checking: - output of /sbin/ifconfig for errors - /var/log/syslog and /var/log/messages for errors - try bringing up a linux machine (not as a thin client) and see can you    consistently mount the nfs share by hand on 100Mb and 1Gb. - ping the server for a while on each port and see is there any difference    in losses (both should presumably be zero). Let us know how you get on, Gavin -- edubuntu-users mailing list <a href="mailto:edubuntu-users@lists.ubuntu.com">edubuntu-users@lists.ubuntu.com </a> Modify settings or unsubscribe at: <a href="https://lists.ubuntu.com/mailman/listinfo/edubuntu-users">https://lists.ubuntu.com/mailman/listinfo/edubuntu-users</a> </blockquote></div>