First let me apologize for joining and posting... but this thing is killing me. I can usually solve these problems myself but I am reaching for help now. I have about 2 years Unix 8 experience but I am by no means an expert but not a newb either.
A little background.
My system runs a particular program, the program I can't really talk about but it runs on unix 8 so that's why I am asking you all for help. I got a call saying my system isn't working and they need a fix. I told them to recover from the backup tapes, they did. Same error. The error is on bootup /etc/hostname.eri0, /etc/hostname.eri1, and /etc/nodename cannot be found and/or are corrupted. The system goes onto login but once the info is entered I get the "DT message cannot be started" and to enter failsafe (which gives the same error) and cycles back to the login screen.
The files are correct, I confirmed this personally. The drive is ok, I swapped them and same error. The drive in question booted up fine on another identical system, which led me to believe it was the processor card. (They are connected by SCSI btw). Swapped the processor card, same error! I am at a loss at what would cause this.
I am sorry for the long post, and I probably left out important information that you would need to make an educated guess. Any ideas?
P.S. - Same error when you login as root. The permissions on the files have not changed and this system has worked for 1.5 years fine now, why the sudden failure??
To be honest I don't know much about that process but I will begin searching. Unfortunately the system in question is around 4000 miles away so I am doing a lot by phone (I was there a few weeks ago). How can that process all of a sudden not start on bootup? Also, this system is very stripped down, I don't ever recall seeing such a process when doing a psit command. It just seems odd that a system that works fine for over a year, with good backup tapes (confirmed) and a good hard drive (confirmed) AND a good processor (confirmed) suddenly shows this error. I even took the system off the network thinking somewhere someone stole the nodename or IP address, no dice.
Still not solved, however I instructed them (on site folk) that an option is to remove the processor and hard drive and put it in a separate VME cage to eliminate the backplane.
ALSO!
They sent me the errors that popup on bootup! About time!
ifconfig: SIOC SLIFNAME for IP: li0
No such interface
It might not be exact, they wrote it on the fly... the lio part... it's gotta be important due to /etc/hostname.li0? The file exists on the disk for sure, but it's not seeing it???
Interesting those are DLPI, i.e. Unix STREAMS, messages. Check the output of "dladm show-link -P". If I recall correctly ToolTalk has a dependency on DLPI.
I removed the fiber interface card from my lab asset processor and recreated the error, almost perfectly. They have a new NIC card on order and hopefully it will work. I will keep you all updated.