[partitionpenguin]

Hey everyone,

Just picked up an Altix 330 on Ebay recently. It was going for somewhat cheap and the listing said it powered on & the fans spun up, though they did not test it any further. Additionally, it seemed as though the Altix 330 didn't require anything else to get booting Linux as a full system so it wouldn't be bad for someone who'd never used an Itanium SGI before.

After getting it at my place and testing it for a little while, I've seemed to run into a little bit of a dead end trying to get any signs of life from it. It was well packed but when I got it the NUMALink connector got absolutely mangled in shipping - the part where the numalink connector was completely bent the chassis. However, it seems like very little actual motherboard damage occurred in shipping, which gave me some confidence. Opening the system revealed nothing obviously wrong and it did actually look fairly clean. Initially I had some trouble using the intended method of connecting a USB NIC to the L1 USB port on the motherboard to get console access to the system, but I circumvented that by using an internal serial port header I found on the motherboard giving me direct access to the L1 console.

Once there, I immediately tried to issue a "* pwr up" command to turn on the system, which did appear to get the system fans running and the system recognized as on (see: http://0x0.st/iHr4.png for env output) but no output from console mode - that is, no EFI shell or anything of the like (When I switched from L1 to console mode using Ctrl-D, I just saw nothing - no output). So I tried digging in the L1 log. This is pretty much the entirety of it.

So, starting around 2009 until 2012 ish is when the system appeared to have been in use (this log went further back to around 2007 when I initially got it, but my repeated attempts to start the system have clogged it a little), and it showed no major errors upon startup, other than occasionally bad voltages. Then for some reason it jumps to 1986. What I presume this is is the system having gone out of service in 2011-2012ish and then the seller attempting to start it up with the clock not set properly. You can see that at this point, it also appears to be working. No major errors. But then, in "2001", which is where my attempts to boot the system start coming in, the system starts spitting out this:

Code:

12/13/01 14:48:09  SH0 MCI LOCK STATUS 0x0 not reached on all L blocks!!
12/13/01 14:48:09  SH0 MCI LOCK STATUS 0x0 not reached on all R blocks!!

I'm not sure what this error is, and Google appears to be no help given the obscurity of these systems, but it definitely seems to be the critical issue preventing the system from booting. It has continued to do this up until right now. I also tried setting the date to something more recent, as you can see, and it is doing the same thing.

So my question to everyone with experience with these systems is - based on the L1 log I attached, what do you think I could do to get it booting, so that I can at least see an EFI p? the clock was dead, which makes me think whatever would be storing environment variables would also be dead, but is there any way to reset the L1 environment variables through software? (I couldn't find any RTC or CMOS battery on the motherboard) Just in case it was set up in a weird way so that it no longer boots now that it isn't set up that way? Or perhaps flashing the ROM, is anyone familiar with doing that on Altix systems? I'm open to trying any fixes you can think of.

[Raion]

We have a few altix boys here. Give 'em time.

[lunatic]

I have no idea on this one. I would probably poke around in L1 and look for more information. The env looks alright. I think you can check some paths, including a console path. What you describe about getting to the L1 console sounds like this might be the problem. Maybe it will get you access to L1, but the system console is set to something else. I had a similar issue with my Prism. When I got it, the EFI console showed up on the graphics like you would expect from a workstation. Then I updated the firmware to install a more modern Linux and the whole thing got reset. The EFI firmware now goes to the serial port like it should for an Altix 350. I actually prefer this new setting, because now I don't have to attach a monitor. You just get another Linux console and X on the monitor if you attach one. But I could never find out where to configure this. It's not in the manual.
The 330 are very rare over here. In my mind, I would treat it like a 350 to diagnose the problem. When you are searching for information, don't be too specific. There is an "adult mode" in the firmware that you can unlock by entering "let the carnage begin", IIRC. Then you can reset a few things that normally shouldn't be tampered with.

[partitionpenguin]

I have no idea on this one. I would probably poke around in L1 and look for more information. The env looks alright. I think you can check some paths, including a console path. What you describe about getting to the L1 console sounds like this might be the problem. Maybe it will get you access to L1, but the system console is set to something else. I had a similar issue with my Prism. When I got it, the EFI console showed up on the graphics like you would expect from a workstation. Then I updated the firmware to install a more modern Linux and the whole thing got reset. The EFI firmware now goes to the serial port like it should for an Altix 350. I actually prefer this new setting, because now I don't have to attach a monitor. You just get another Linux console and X on the monitor if you attach one. But I could never find out where to configure this. It's not in the manual.

The 330 are very rare over here. In my mind, I would treat it like a 350 to diagnose the problem. When you are searching for information, don't be too specific. There is an "adult mode" in the firmware that you can unlock by entering "let the carnage begin", IIRC. Then you can reset a few things that normally shouldn't be tampered with.

Huh. Thanks for the info. I might try that "adult mode" if I can find anything about it. I actually completely forgot I made this post. I ended up calling it quits and finding an Altix 350 on ebay (from the same seller, as it turns out), as well as a source for a PCI card cage. That altix is now happily running SLES, which is really great for me.

But I decided to come back here because I really wanted to leverage the built in CDROM in the 330, so that I could try booting operating systems that don't play so well with netboot (what I used to get SLES on the Altix 350). When I looked back at the logs, it occurred to me that one of the other messages printed out in the console suggests that it's looking for some kind of numalink router/something to connect to (I'm not really familiar with numalink), and a slot in the rack.

It would be great if I had some way to reset the system (like clear a PROM). I think this might be able to clear any numalink configuration it has. Alternatively, now that I have an Altix350, I could try booting it with a NUMALink cable connected, but I'm not sure if that's worth it. They're not terribly expensive on ebay, but it could just be a waste.

So I'm still looking for suggestions for how to clear environment variables, reset the PROM, or whatever the equivalent may be on Altix systems.

[lunatic]

As far as I remember the Altix 330 cannot directly link with the 350. 350s can be linked just by connecting them with the cables in the correct sequence. 330s needs a Numalink router.

On the other hand, a machine has to set its rack slot before it can join a numalink configuration, because the rack slot is the node identification. It really seems like the L1 firmware on your 330 is seriously broken.