Proliant 800 POST problem

[Hugo Teerds]

Hi Guys.

Since last week my Proliant 800 has POST problems. At first it told me '101-I/O ROM error' for some days, and after reflashing the BIOS it worked every time.

Now since three days it does not start at all, it stops before teh COMPAQ logo appears. I then have to do an emergency flash by setting the dip switches every time. After that it runs as usual, but the next day all is lost again.

I know from this forum that when the box loses the date and time, it is time to put the replacement battery in. Can it be that this problems is also caused by the battery end-of-life? I would like to have some opinion before I rush out and buy a new one.

[JohnWRuffo]

Hugo:

No... the batttery should not be causing the bios to loose the ROM... it is burned in. If I were to put my money on the problem I would say the ROM chip is shorted. This is a M/B problem.

A dead battery would be a "System Options Not Set" POST error. This is alltogether a different animal.

Have you tried to flash the ROM with a prior version and then re-flash back up? Try a different diskette?

I hate to think this is a main board failure but I am inclined to say this is.

-john

[Hugo Teerds]

Thank you for the quick response John.

After the emergency flashing, the server has to be power-cycled and the dip switches set back again. This takes less then half a minute, but during that time the power is also off. After restarting all is well.

It seems for now that things go wrong when the power is off for a longer time, such as a night.
Is this consistent with your vision of a shorted ROM? All I have done when the problem started was disconnect all RAID-adaptors and I/O boards once to check for the reason. Could that have done it?
And the last one for now: Is this particular chip replacable or has the entire motherboard to be replaced?

Thanks for you trouble, I am a faithful reader of these forums and see what effort is put into it.

[Hugo Teerds]

Oops, I forgot to reply to your suggestions.

I will make new disks with the former and latest BOIS's tomorrow and try them.

I will let you know the results.

[JohnWRuffo]

Well... here is what I am working from. Take a read of:
http://h200001.www2.hp.com/bc/docs/support/UCR/SupportManual/TPM_186861-001_ch3/TPM_186861-001_ch3.pdf
Look on page 3.

Now, from my experience, a 101 ROM Error can be caused by a bad flash (but would normally POST with a CRC error) ... but primarily this is caused by a bad ROM chip.

If you can get this to POST doing some magic and re-flashing it, I would suggest you immediately run a full diagnostics on the system. A quick pass first and then a full loop testing everything. If the quick pass finds an error, I am sure the error is related to a hardware malfunction.

Here is the diags I am thinking of:
http://h18023.www1.hp.com/support/files/server/us/download/18396.html

Let me know what you find?

-john

[Hugo Teerds]

Thanks John, I will.
It is getting a bit late here in Europe, so I will quit for now. I will tell you tomorrow what I find.

[JohnWRuffo]

I will likely drop by the boards tomarrow and read what you find.


what am I saying... I am always here. See ya then!

:-)

[JohnWRuffo · Accepted solution]

Hugo:

After you reflash the ROM, you say it will boot? Have you re-run the System Configuration Utility and saved the config? If not, please do this prior to running the diagnostics. Lets make sure the system knows about itself before the inspect tests.

http://h18023.www1.hp.com/support/files/server/us/download/13227.html

Let me know what you find. Let me know if any of this has been helpful?
-john

[Hugo Teerds]

Hi John,
Glad to report that all is normal again.
It turned out to be my mistake (of course!)

What has happened is that when I first got the 101 errors last week I used the SCU and saved the new configuration. And after that the box would not POST anymore. I forgot that that happened before when I added a small USB PCI card in the system last year. The only way that I could make the system work normal again was take the board out, do a configuration and safe it, and then add the board without new configuration. Not a nice or recommendable practice, but for this board it worked without problem.

And you guessed it, my new configuration was made with the erronious board in place. (shame on me for my bad memory!)

So I now remembered after your previous message, configured teh system again in the wrong (but for me here right) way, and all was running again. To make sure I flashed the BIOS again and tested with the diagnostics disk. There now seems to be no more problem at all.

Thank you very much for the help in thinking and realising my erroneous ways.

[Hugo Teerds]

Update:

In the logs there still were some 101 errors although they didn't show anymore during POST. And sometimes during startup the server would beep three times, one long and two short beeps. The manul tells this is due to a display adaptor failure.

So I replaced my Matrox Millenium G200 board (Compaq nr 402543-001) with another one I had still lying around, and am happy to report that that finally solved all problems.

I thought I would tell you, as I can't stand unresolved cases myself. I hope someone once will be helped with this thread, as I often am with the problems of others here.

Thanks for all help and efforts.