health led blinking red

[RNeville]

hi guys,

i have a hp160g6 server runing centos as a webmail. everything works fine since november 2009 until about 2 months ago.

now the server hangs unexpectly and shows the front health led flashing red. no errors in logs, the maillog shows the server working normaly for a few hours and then nothing...

ssh shows the login but hangs after i enter the password. the same happens locally. the only opcion is restarting from the power button.

hp support recomends to turn off graphic mode and a bios upgrade, i did both things and nothing. they didnt find any problem with the hardware or system.

sometimes i got this error:

Uhhuh. NMI received for unknown reason 30.
Dazed and confused, but trying to continue
Do you have a strange power saving mode enabled?

Uhhuh. NMI received for unknown reason 20.
Dazed and confused, but trying to continue
Do you have a strange power saving mode enabled?

googling i found that disabling hpet (nohpet as kernel option) can help with this problem, i tried but now hangs with no flashing health led. i tried too nmi_watchdog=0 with the same results.

im using 2.6.18-194.3.1.el5 but the problem appears with 164.11.1 or 164.15.1.

anybody has an idea for fixing this??


thanx in advance

pd: sorry for my poor english ;)

[Matti_Kurkela]

You probably mean DL160 G6.

NMI means your hardware is trying to tell you something, but the "unknown reason 30" means your Linux currently does not have the means to interpret what it says.

You might want to install at least the "HP System Health Application and Command Line Utilities for Red Hat Enterprise Linux 5" RPM. The name of the actual RPM should be something like "hp-health-.rpm"

It is available in http://www.hp.com/go/support (the standard download location). Select "Download drivers and software" and type in your server model. On the next pages, confirm the server model and your OS, then you'll see all the downloads for your OS version.

If you're using CentOS 5, it will probably be possible to download the RPMs intended for RedHat Enterprise Linux 5 and install them. (you may have to edit /etc/redhat-release to say RedHat instead of CentOS if the package checks it, though)

After the RPM is installed, run "hpasmcli", then type "help" and "help show". The output of hpasmcli commands like "show server", "show iml", "show dimm" might shed some light to the problem...

MK

[Tim Nelson]

Information could also be gleaned from logging into the ILO and reviewing the event logs.

[RNeville]

hi matti,

after all thanx for you reply.

i understand the issue with the nmi problem message and the system. when i saw it the first thing i done was install the hp-health rpm, but when i contact support for the first time they told me to uninstall every redhat package from the system because compatibility issues(?)

i will try installing again but the values was right then, except for "show iml" that returns not present, not activated or something like that.

i have no hp packages installed at all actually.


thanx again...