Question regarding BL465c G6

[chuckk281]

Patrick had a customer question:

********************

Hi again everybody.

We are always working with this issue;

For you information, the ALtiris Agent was stopped an all the machines, but the problem is still here..

We are trying the capture “Kernel Dump” during the Freeze.

We have configured most of the machines with registry key NMiCrashDump and it seems all the prerequisites for generating a Kernel dump are fulfilled.

As the problem is really short in time (1-2 minutes), customer is not sure to be able to generate the dump at the right moment.

Do you think it is possible to  script the click on button  “generate NMI to system”…?

The idea is to test the ping echo replies from the faulty machine and to launch the dump when sufficient amount of “Request Timed out” replies are received from this machine.

*****************

Info from John:

******************

If you want to script the dumping of a system you should use the NOTMYFAULT utility from Sysinternals as discussed here to trigger the dump:

http://support.microsoft.com/kb/969028

The question is how do you work out when to run this command?   Does anyone know if you can use PROCDUMP to trigger the use of another executable, ie., PROCDUMP detects a particular PerfMon counter has hit a certain limit using its “-p” option and then calls NOTMYFAULT to trigger a memory dump. 

Don’t forget that capturing information from XPERF may also be useful here as I had previously mentioned to you.

**************

Other comments or suggestions?