topic Re: Problems with console and temperature in Operating System - HP-UX

Problems with console and temperature

Pedro José — Thu, 01 Apr 2004 02:56:45 GMT

Hi everybody!!
We have just installed HP-UX 11.11 in a new server machine.
At some moment on the boot proccess it seems to be hang (it does not show the login prompt) and after some time, the next message is shown on the console.

********SYSTEM ALERT********

SYSTEM NAME: uninitialized
DATE: 03/23/2004 TIME: 21:50:13
ALERT LEVEL: 5 = Boot possible, enviromental problem exists

REASON FOR ALERT
SOURCE: 6 = platform
SOURCE DETAIL: 1 = cabinet SOURCE ID 0
PROBLEM DETAIL: 1 = inlet overtemp

LEDs: RUN ATTENTION FAULT REMOTE POWER
ON FLASH OFF ON ON

LED State: There was a system interruption that did not take the system down.
Check chassis and console logs for error messages.

0x00E000516100E000 00000000 00000000 - type 0 = Data Field Unused
0x58E008516100E000 00006802 1E15320D - type 11 = Timestamp 03/23/2004 21:50:13

I have read some articles about this error and itÂ´s something related to a high temperature on the server, but the room in which the server is located is about 24 ÂºC.
I verify the error and exit from the GSP but the login prompt isnÂ´t shown anyway.

Then I have two questions...
- What is happening with the "inlet overtemp"?
- What is happening with the login prompt?

Thanks in advan

Re: Problems with console and temperature

Chris Wilshaw — Thu, 01 Apr 2004 03:05:55 GMT

It's possible that one of the internal fans on the server has failed, which is causing one part of the system to overheat.

As the temperature in the room doesn't seem too high, I'd suggest logging with HP.

Re: Problems with console and temperature

Camel_1 — Thu, 01 Apr 2004 14:04:37 GMT

If you checked all fans are working properly, it might caused by high temperature in the server room.

when everything alright you can reset the error in GSP mode.

Simon

Re: Problems with console and temperature

Chris Vail — Thu, 01 Apr 2004 14:58:59 GMT

For less than $100, any good electronics store will sell you a laser thermometer. A company (such as Grainger) that wholesales air conditioning and heating supplies also sells a airflow gauge. Those two devices will pretty quickly help you deterimine if the envirnoment in your data center is appropriate.

You need to check the temperature of the air going into the system, to compare with the temperature of the air coming out of it. HP (and most manufacturers) publish the number of BTU's each machine should use.

Although your data center averages 24 degrees C, the area immediately near this machine could be significantly higher. I've seen temperature differences of 5 degrees C between adjacent racks. In those cases, a regular room-sized fan is needed to move the air around and mix it more thoroughly. Until you check the temperature of the actual air input of the device, relying on a thermostat in one part of the data center is just not good enough.

The same is true with airflow. Check the volume of air going in, and going out. They should be the same except for very small amounts (due to leakage in the cabinets).

On large systems, such RP8400's and Superdomes, HP offers a free preliminary site survey. I strongly urge you to take advantage of these whenever possible.

Until you get the hardware I mentioned, I recommend you open a call with HP support for the answer to your overtemp question.

Chris

Re: Problems with console and temperature

Pedro José — Fri, 02 Apr 2004 01:19:35 GMT

Many thanks for the responses!!
IÂ´ll contact HP support for the overtemp problem to look for all the fans working properly.
Now I repeat the question of the console login. As I explain in my original post, the login prompt doesnÂ´t appear at the console, but I think all the proccess are working OK. For example, I can make a telnet from other machine on the net without problems and look for all the proccess on the system.
From the console (empty screen) I enter in the GSP, have a look to the error messages (overtemp) and exit (CO) but the login prompt isnÂ´t still at the console.

May it be cause of the overtemp message?
If it isnÂ´t, which can the problem be?

Thanks

Re: Problems with console and temperature

Naveej.K.A — Fri, 02 Apr 2004 01:46:49 GMT

Hi,

Just check once again for any error saying fan failure or so...at a bit earlier time than the Over temp notification.
Normally GSP gives a fan failure error mentioning the fan number incase of a fan failure.
This is clearly due to temperatue in the room which has increased above the threshold level..

Login to the server from any of the other client using telnet and try out fuser -ku /dev/console and see whether the login prompt is coming back!!

With best wishes
Naveej

Re: Problems with console and temperature

Shaikh Imran — Fri, 02 Apr 2004 01:54:43 GMT

Hi,
what is the Model of your system
is it A class i.e rp24xx ?
Reg

Re: Problems with console and temperature

Pedro José — Fri, 02 Apr 2004 02:41:45 GMT

Hi Shaikh!!
My system is rp 2430.

Re: Problems with console and temperature

Shaikh Imran — Fri, 02 Apr 2004 04:01:19 GMT

Hi,
There is a known Problem with rp24xx
The speed of the CPU fan reduces suddenly as said by HP. for that Hp is adding a small controller/conductor in series with the CPU-FAN and the CPU-Board.
Hey Pls contact HP for the same
Last week only we had done the sam with our
rp2470.. 2Nos.
You will require one for each CPU if you have SMP.
you will also have to get the downtime for the same which is Nothing but only 1/2 and hour.

Regards,