[SterlingNOVA]

My situation:

Have six Tezro rack sets I'm doing repairs on. Processor and Video bricks
configured for dual display. Irix 64 6.5
All were being finished up, system tested and all checked good.

Came back next morning and was setting up to do final checks.
First set I powered on, system booted to dual display SGI logo splash screen and hung.
I then received a "DT messaging could not be started". Have had this happen before, I got a second
drive and restarted. The unit booted up fully and all hardware checks were good except the console out
check. Entered  "cu -s 9600 -l /dev/ttyd3". It returned an error message "Connection Failed. No Devices Available."
Originally thought I may have a bad port on this repair, however, using the working drive I had I booted the other systems
to check them. ALL SIX sets were failing for same error. I'm certain that this is a software/corrupted disk issue, and not hardware.

As stated before, all the sets had previously checked good.

Have a good console port in to the L1 controller. Cannot find any errors. From system startup menu checked enviroment setting, inventory (hinv)
nothing appears out of whack.

I'm not anywhere close to being an expert in IRIX. Anyone have suggestion where to look in the filesystem to straighten this out?
Appreciate any help.

[Raion]

Does this happen on boot into the PROM or into the operating system?

[weblacky]

I'm not an expert on this...but a google search for your term "DT messaging could not be started" brought up old UNIX forums where the discussion was loss of a station's previous network settings when relying on NFS (or the like) mounts for user home directories or other system-required directories. During this error, only root could login as root wasn't assigned a network mounted homedir.

So not knowing much more, can I assume you rechecked that networks settings are still what you remember and your DNS/hosts file info is in order? Maybe it's as simple as something lost it's network-related config on shutdown and on reboot was assigned new settings that cannot access the remote server required for a process to start (thus halting the system).

Other than that, we'd need much more information to help you.

Aside from that, I'm personally curious, what are these stations being utilized for? Seeing as you don't hear about SGI being setup for use much these days.

Thanks,

[hamei]

The unit booted up fully and all hardware checks were good except the console out
check. Entered  "cu -s 9600 -l /dev/ttyd3". It returned an error message "Connection Failed. No Devices Available."

The console out check ? Usually you use the console to get in, not get out ... and a hardware check on console out ? I could easily have missed that since I never look for one, but don't remember ever seeing that.

Just for fun, have you tried booting from the console ? There's more messages that go past, especially if you setenv v ...

from the prom via a terminal (ctrl D), first #5 command line then boot -f /dev/dksc(0,1,8)/sash  (I think you can just boot -f sash too)

then from sash

boot -f /dev/dksc(0,1,0)/unix

and you can watch more stuff scroll by.

But the rest of Irix is working ? I would not worry about it ....

You can do an < ls dksc(0,1,0) > also, to look at what is on the boot disk.

[SterlingNOVA]

Does this happen on boot into the PROM or into the operating system?

The failure is occurring in the operating system, when checking the serial port (My bad, not console).
Opened a terminal, and invoked the command to established a connection to a network switch. (cu -s 9600 -l /dev/ttyd3)
All other hardware functions are fine, including the port into the L1 controller.
Like I mentioned, all six of these sets worked normally on previous tests. Now the problem is across all six sets.
It is almost as if the ports are disabled somehow in the software.

[hamei]

The failure is occurring in the operating system, when checking the serial port (My bad, not console).
Opened a terminal, and invoked the command to established a connection to a network switch. (cu -s 9600 -l /dev/ttyd3)

That makes more sense. You tried all the serial ports ? Have you tried the confidence tests ? Stick a serial device on a port, boot up and see if it gets seen ?

Just out of curiosity, are these six single-brick Tezros or are they doubled up ?

[indy99]

Try switching your serial communication speed to 38400 rather than 9600.

[hamei]

Try switching your serial communication speed to 38400 rather than 9600.

Not to be a Debbie Downer but I don't expect that to make any difference. Lots of serial devices run at slower speeds, you are setting the speed from inside, not connecting to it from outside.

But for the op's sake I hope I am wrong.

But I'd bet fifty cents there is no problem, just something OP is doing wrong.

[SterlingNOVA]

Baud rate isn't the issue, always used 9600 for the test.
Ran across some info about running cu in debug mode, so going to give that a shot and hope it points to something missing.

To the DT Messaging Error. Appreciate the input and I have have been down the trail of going through host files and address
checks. It has never resolved that particular issue. For whatever reason, the DT error is usually preceded by some panic
alert and dump to the disk. The only resolution I have had is re-imaging or cloning from a master disk.

Which is one of the strange things about this. The disk I'm using is a previous clone that was working perfectly fine.
It was not installed when the DT error occurred, and should be a clean error free drive.

[weblacky]

Ok, rereading your initial description...is it at all possible your “sets” are using NumaLink? Again, I only own desktop Tezros (not rackmount). But could this all stem from the fact that you’re not powering the “sets” up in the correct order? Or someone fiddled with the rear cabling?

In my mind, if you boot from another disk that wasn’t NUMALink configured then everything is fine, if you later install software and settings that activate NumaLink functionality, then missing its connected twin would definitely cause an error early on.

The manual doesn’t say what happens if you leave a NUNAlink pair this way.

Could you recheck the rear of the installation and ensure your powering either everything up at once or that if you starting in staggered sequence that you are powering the NumaLinked pairs together and not accidentally powering on one of multiple sets?

From what I know, Tezros can only NUMAlink as one pair, so these “sets“ you refer to struck me as odd word choice. Perhaps this is why and the simple answer to the unexplained behaviors?