3 Silver

Re: Random Reboots with a R710 and Windows 2008 R2

I am being told that its one or the other, disabling c states and hotifx together shouldnt be necessary, are you seeing otherwise?

0 Kudos
1 Nickel

Re: Random Reboots with a R710 and Windows 2008 R2

You are essentially correct insofar as I have not tested that configuration.

I would do the following steps in this order:

1) Disable C1E.

2) Install Hotfix.

3) Disable all C States.

Although I've done all of the above I would go in order of 1, 2 and then do 3 if 1 & 2 are not fixing the problem.

0 Kudos

Re: Random Reboots with a R710 and Windows 2008 R2

There are several issues going on in this thread vs a common issue, imho - some are getting blue screens and some are not (some just reboots) and there is little info to link the issues other than a "reboot", so I'd prefer to be cautious and consider each issue as a separate issue.

1) For Hyper-V, install KB975530 - this fixes issues with Nehalem procs that are seen when C-states are enabled in BIOS (this is not C1E, but all C-states).  If this KB is installed, C-states can remain enabled.

In Dell BIOS, under the CPU category, there is a "C-State" enable/disable field - this is the field that I am referring to. It looks like there is also a separate field for C1E - but this would likely not impact the watchdog timeout bluescreen fixed by KB975530 - all C-states must be disabled OR the KB must be installed and c-states can remain enabled.

2)  The other KB mentioned in this thread - KB974090 - is associated with C1E, but is not, to my knowledge, associated with any blue screens - it simply improves C1E usage.  That can also be installed, if you wish.

3)  WRT blue screens - without knowing exactly what the blue screen message is, I can't comment - can I get the exact blue screen message?  Since it reboots upon blue screen, if you want to halt the system on blue screen to read the error, you can do the following:

1)   Open a command prompt

2)   Run regedit

3)   Go to HKLM\SYSTEM\CurrentControlSet\Control\CrashControl

a.    Modify the Autoreboot setting from 1 to 0

b.    Reboot the system after modifying the registry, then retest the scenario where the failure occurs.

c.    The OS will then blue screen, but remain at the blue screen so we can read the blue screen error message

You then have to  reboot the system manually once you get that info.  In addition to getting the blue screen info, you may also see if it writes a memory.dmp file (located in c:\windows) - it will show on the blue screen if it wrote a dump file.  This can also be used to get much more information on the exact root-cause.

0 Kudos