ML370 G3 - how to generate NMI dump?

[Geoffrey Whall]

I need to generate a NMI dump when my server becomes unresponsive. The generate NMI button on the ilo doesn't work and I cant seem to locate a button on the MB which might do this.
any ideas?

[KarloChacon · Accepted solution]

hi

location in SB, page 93 or 4-14

http://bizsupport.austin.hp.com/bc/docs/support/SupportManual/c00064921/c00064921.pdf

also check this
for this you can assign Ilo hot keys for example Control+T to CTRL key and the SCROLL LOCK two times.

Control-T = R_CTRL SCRL_LCK SCRL_LCK

http://support.microsoft.com/kb/244139

regards

[Geoffrey Whall]

thanks
that's the button I was looking for, but it doesn't generate the dump when this system freezes. I've also tried the MS article, just cant get it to dump, I've replaced every piece of hardware in the system so I'm pretty sure its not a hardware problem causing it to freeze. MS are telling me they cant do much of anything without a the dump file.

[DSW_2]

I have two ML370 that exhibit the same system lockup. No messages or dumps. Windows 2000. No blue screen on system lockup. System freezes in the middle of painting the normal screen contents. No dump or event log entries available after cold reboot.

Normal server operation (Oracle RAC) runs without lockups. Lockup only occurs with users logged in on the system console and using the GUI.

HP has replace all components in one of the units. We have been unable to find a solution for this problem.

[juan quesada]

hi,

in the RBSU (BIOS) under advanced option, is the NMI button enabled?

regards,

[Blazhev_1]

Hi,

here is the official documentation how to generate NMI, check prerequisites and step-by-step guide :

http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00797875/c00797875.pdf

There is an option in iLO for virtual NMI(to generate it through ilo, not physically pressing the NMI button).

However, not sure if this is the best action...

You replaced the whole server and same situation?

Lockups without bluescreen are mainly caused by old firmware(BIOS, controller and HDD). If HP replaced all parts ,you definitely should have done this, but just in case :

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=15351&prodSeriesId=316541&prodNameId=3279715&swEnvOID=1005&swLang=8&taskId=135&swItem=MTX-4fc853c939da4139af2a1217fc&mode=3

You have also replaced cable between controller and backplane?
Are there any errors in IML or absolutely nothing?

Regards

[Geoffrey Whall]

hi

yep everything has been replaced, all firmware has been updated. We even moved the disks to another server same thing happened so we are pretty sure its OS related, W2003 in my case.
No errors logs anywhere, ILO NMI button doesn't work as well.

our workaround is to schedule a reboot of the server twice a day 8am and 8pm, for us this has stopped the lock ups. We are going to rebuild it from scratch when we get some time.

[Graeme Bray]

http://support.microsoft.com/default.aspx?scid=kb;EN-US;927069

Try following these steps. I was able to successfully generate a NMI dump on a Win2k3 box using this method.

This was also what a Microsoft Tech had sent me.

3. MEMORY DUMP INSTRUCTIONS:



1st make sure that the machine is setup to get a complete memory dump:

1) In Control Panel...double click the System Applet

2) Select the Advanced tab

3) Click the "Startup and Recovery" button

4) Under the "Write Debugging Information" section select:

"Complete Memory Dump" from the pull down menu

Note: If â complete memory dumpâ is not an option found in the dropdown menu, we can configure it via registry.

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\CrashControl

Make the value of â CrashDumpEnabledâ to be 1.

5) Make sure a check mark is placed on:

"Overwrite any existing file"

6) Make sure that there is a paging file (pagefile.sys) on the

System Drive and that it is at least RAM + (.5 x RAM).

7) Also make sure there is more space on the hard drive than there is physical RAM



Note: There is a limitation with the Kbdhid.sys driver that allows for you to generate the memory dump process by using a USB keyboard. The CTRL+SCROLL LOCK+SCROLL LOCK keyboard

shortcut does not work if the computer stops responding at a high interrupt request level (IRQL). This limitation exists because the Kbdhid.sys driver operates at a lower IRQL than the

i8042prt.sys driver. Please install the hotfix in KB244139 to take a dump using the USB keyboard.



===================================



Second, configure the system to dump memory:

8) Start Registry Editor (Regedit.exe).

9) Locate the following key in the registry:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\i8042prt\Parameters

10) On the Edit menu, click Add Value, and then add the following registry value:

Value Name: CrashOnCtrlScroll

Data Type: REG_DWORD

Value: 1

11) Quit Registry Editor....

12) Then reboot the machine.

========================================================================

When the machine exhibits the problem do the following keystrokes to generate a memory dump:

Hold down the RIGHT CTRL key, and press the RIGHT SCROLL LOCK key twice.

========================================================================

[Geoffrey Whall]

Yep, tried that as well, that was the first thing MS said to do

no luck

[Graeme Bray]

And you tried to update the RILOE card's firmware also?