Issue in troubleshooting an ancient HP9000 K260, running HP UX 11.11

[Patrick Scalia]

I'm a technician who has been called upon to troubleshoot an ancient HP9000 K260, running HP UX 11.11.

After a power sag/crash/restart FSCK gave multiple error/warning messages about logical volumes and groups. I have some modest experience with Unix servers, but not on such an old and complex machine, and I'm in way over my head. Before I even go into the problems I'm asking for help with, is there anyone on these forums who has knowledge of such an old machine? I would be extremely grateful for any help I could get. I suspect file system corruption and perhaps a failed physical drive, but I don't even know how to figure out which drive (of the 14) needs to be replaced! I have replacement drives, and happily data loss is not an issue because there was a complete backup of the OS (Ignite) as well as all data.

Yes, I know this machine should have been retired years ago, but for whatever reason the client is willing to spend the money to keep it going for a couple more years. There seems to be nobody in my entire state who has experience working on this thing, however. Including me.

I'm accessing the server through an equally-ancient terminal, not a PC, so I can't copy and paste output here on the forum. I do have clear .jpg images of the screen output, and can attach those.

I will be extremely grateful for any and all assistance.

[Dennis Handly]

>I'm accessing the server through an equally-ancient terminal, so I can't copy and paste output

Was it a fancy HP terminal? You can read the memory of those to copy the screen.

[Patrick Scalia]

Thanks for the reply.

It's an HP terminal. Whether or not it could be considered "fancy" I don't know. I haven't used terminals much so I don't have anything to compare it to. I don't understand what you mean about reading the memory to copy the screen. If it's necessary I'll type it all out by hand to post it here, if it will help me get assistance. I have a lot of pressure on me to find a solution. Once again, I'll be grateful for any help I can get.

[Bill Hassell]

The K-class may seem to be ancient hardware but in HP terms, it's just a bit past middle age. The fact that you are running a supported version of HP-UX (11.11) means that you'll get a lot of help here in the forums. HP terminals are quite smart (compared to a vt100) but that doesn't help with your troubleshooting -- it's a serial terminal.

So the first thing to do is to get a laptop with a serial port (now that is ancient hardware these days) or at least a working USB-to-serial converter. Now you can capture the console messages using PuTTY or a similar terminal emulator. NOTE: The serial cable for the laptop will be a 25pin to 9pin printer cable (also called a crossover cable). Once you have this in place, you can start capturing the information.

Note that if this forum is the only resource you have to get the hardware fixed, you may be down for several days or weeks. If you need knowledgeable help getting this system up and running, I can get some names of service companies that can help -- send a note to bill dot hpux at gmail dot com.

And as you have seen this machine (and probably several others) are at risk due to the lack of a smart UPS to protect equipment from short and long term power failures. Downtime costs should be evaluated and compared to the cost of a properly designed smart UPS system.

[Patrick Scalia]

I guess the lesson in all this is not to bite off more than you can chew. We're going to have someone with actual experience on that machine come out to look at it. If I'd known before how to get in touch with him, or someone like him, we'd have done that from the beginning. All HP had to say about it is "We can offer no support whatsoever for that machine." Many thanks to Bill Hassell for putting me in touch with the guy.

[Michael Steele_2]

Refer to the front panel LED especially during boot up. Any errors will be displayed in this panal.

There is no GSP or MP on this box, however, the last 50 LED displayed codes are kept in log files.

Reboot the box and copy down any errors from the LED. They'll cross reference to some book that you're supposed to have.

[Michael Steele_2]

Front PAnel Error Code Meanings for K class Sever.

0 Reserved (not used)
1
Interrupts
Selftests, Diagnostics
Boot Codes
2 Cache (selftests and diagnostics)
3 Processor Dependent Hardware (PDH)
4 Late Selftests
5 Bus Transactions
6 Reserved
7 Memory Subsystem Faults and PDCE_HPMC Memory Faults
8 I/O Device Faults (K-Class/D-Class)
9 Console Initialization Errors
A Boot Device Initialization Errors
B Operating System Panic (OS)
C
System Initialization
HP-UX ISL Boot Codes
HP-UX System Initialization Codes
D Shutdown Codes (OS)
E Warning Codes (OS) (not used)
F Run Codes (OS)

[boukari]

Hi ,
try to check syslog file and eventlog, then you can use this mail script to send them to you and well you can share it with us:

/var/adm/syslog/sys.log.log
and
/var/opt/resmon/log/event.log

Mail script
http://www.hpux.ws/?p=7

Hope that helps,

[Patrick Scalia]

Well, the front LCD panel never displayed any error messages at all. Which is why I assumed that if there was a problem with a physical drive, that the drive was probably in one of the two attached drive enclosures rather than in the main body of the server itself. Again, just assumption, but I would think the LCD would only display messages relevant to the server and not to an attached peripheral, which the drive enclosures are. Of course there's a good chance I'm wrong about that. Anyway, thanks for the input.

[Dennis Handly]

>Whether or not it could be considered "fancy" I don't know.

Fancy compared to a tty or even a glass tty.

>I don't understand what you mean about reading the memory to copy the screen.

I had a pscreen program that would copy screen memory to a file or printer. But once I had X windows, I used cut&paste.
A jpg would be a good first start.

>but I would think the LCD would only display messages relevant to the server and not to an attached peripheral

Yes. But it may say the system halted or crashed, not particularly helpful.

[Michael Steele_2]

"..Well, the front LCD panel never displayed any error messages at all..."

Before the drive is called many, many messages are going to be displayed. For example, cpu test, dimm test, etc. Drives aren't going to be involved until all the firmware has been verified and its time to load the O/S.

Google boot sequence / order for hpux for conformation.

So, that you see NOTHING!

NOTHING! AT ALL on the front panel????

Then something basic like the key on the front panel is in the wrong position or the power is unplugged, or a cable has been dislodged.