I have a Sco Unixware 7.1.0 server which has been running fine for nearly 10 years ... we have not had to do any os changes for several years.
We tend to reboot it about once a month just to keep everthing clean!
My issue is that we have just rebooted it and on startup the network is not now working - I cannot ping anything other than the loopback or the server itself.
I have replaced the NICs (for the same model Intel Pro/100B), changed slots in the server and even restores the entire system from tape .. all to no avail!
The strange thing is though that the netcfg command seems to hang when trying to reconfigure the network - I have managed to remove the nic and re-configure it after severasl attempts but even when I have the network has not come back.
I really am lost now so any help would be appreciated ..
Another strange thing .. I restored the OS disk off tape and rebooted - the network came back up .. when I shut it down and put the server back in position and then rebooted - no network again!
Restored from the same tape as before, but this time no network!
I know the NIC is ok as I replaced it from a working server.
The NIC is being seen by the OS as if I unplug it from the switch I get the eeE Lost Link message on the console .. I just cannot get it to talk to the network.
As I say the worrying thing to me is that running netcfg
a) takes a long time to pick up settings
b) when it does come back I cannot see all the options (greyed out) in particular, modify the settings
c) once in netcfg, I cannot then exit .. it justs stays within the menu system.
Again very strange that I can restore the system from a known working date but still cannot get the network running!!!
Lets suppose that some file corruption/misconfiguration, or whatever you want to call it occurred, and didn't manifest itself until the computer was rebooted, can you be sure that the corruption is not also on the backup tape.
Can you test your oldest backup?
What about trying a different brand of NIC.
How long can you be down for before replacing/re-installing the system becomes a viable alternative?
I had a similar issue with 7.1 but it was always with changing the IP or gateway. When making a change in 'netcfg', not all files would get updated with the new config, so it caused a conflict.
I had problems so often that I saved the tech article on SCO's site:
Basically, I had to do a search on all files which contained the old IP address or the old gateway address, output the names of each file to a list and then edit each file listed with the new 'correct' IP/GW.
I had so many problems with UW 7.1 over the years, but since upgrading to 7.4 in 2005 I haven't had a single issue of any kind.
Sorry for overlooking the first thing that came to mind, but I assume you've ensured that you have a default route added somewhere in a startup script in /etc/rc2.d ??
For instance:
route add default <gateway ip address>
Just making sure. Sometimes it's the obvious stuff that gets overlooked.