I can't get beta3 to install over NFS. Beta2 works fine with exactly the same input: - no special options at syslinux prompt - English language for installation - UK keyboard layout - NFS image installation medium - NE2000 driver io=0x300 - IP address 192.168.1.5, netmask 255.255.255.0, default gateway 192.168.1.1, primary nameserver 192.168.1.1 - NFS server name "cyberelk", Red Hat directory "/usr/local/src/big/cdrom" With beta2, it just works. With beta3, the directory gets mounted (cyberelk authenticates the mount request), but then a dialog box says: I could not mount that directory from the server VT3 says: ... mounting nfs path cyberelk:/usr/local/src/big/cdrom VT4 says: ... <4>eth0: NE2000 found at 0x300, using IRQ 5.
Try specifying the FQDN for the NFS server. Sounds like something is not connecting with this install. We are having no problems performing NFS installations from porkchop, so I am pretty amazed that you are.
Created attachment 101 [details] Output of tcpdump -ieth0 -x -s 500
Well, I am. I tried using the FQDN but it made no difference. I tried using a US keymap too, just in case. ;-) With 6.1.91 against the same NFS server (Red Hat Linux 6.0) and directory, it gets further, i.e. to the point where it sees a different tree to the one it's expecting. I caught the network traffic -- see above.
I couldn't reproduce this on the machine here in the office that has an ISA NE2000. :-( On my home machine, this happens _every_ time, and I don't even know a workaround.
The way that I am reading this, you are performing an NFS installation from your house and it is failing. If this is the case, please have a conversation with Bryan Andregg about global access to internal redhat machines. He will probably be able to shed light on your problem.
No, I'm using two machines at home, one of which is the server, and one the test install box. But I can't reproduce this on any other machine, so I guess I'll leave this closed.