UPDATE - May 25, 2010 - It appears that our Dell DSET problem is hardware-related. We seem to have some hardware problems which are causing the DSET utility to reboot our ESX host. If you are having this issue, you should call Dell and let them know (you probably are already in contact with Dell if you're running DSET).
We'll let you know what we find out with regard to our hardware.
---------------------------------------------------
We've had an ESX host that has been purple screening (Puprle Screen Of Death - the ESX equivalent of a Window Blue Screen) and it appears that a bad CPU is causing the issues.
In order for Dell to get the logs necessary to dispatch a new CPU they requested that we run the Dell DSET utility - which is supposed to just gather log information and package it all up nice and tidy so I can send it off to Dell.
When running the utility there is no mention of a reboot, or possible reboot that could be necessary to gather the requisite information. However, when I ran the utility I heard a 'beep' from my terminal and low and behold I see the following on my screen:
Since the report file path is not specified, the report will be created in user home
The options selected for the report are: --time --advanced -f /root/
Dell System E-Support Tool
@Copyright Dell Inc. 2004-2010 Version 1.9 build 131
* Getting Linux system summary information ...
Gathering Network Information ...
Gathering OS Summary Information ...
* Xsltproc is not installed. Installing now...
* Uninstalling the installed xsltproc..
* Getting Linux operating system configuration information ...
Gathering Boot Information ...
Gathering Module Information ...
Gathering Memory Information ...
Gathering Storage Information ...
Gathering Network Information
Gathering Summary Information ...
Broadcast message from root (Mon May 17 15:07:21 2010):
The system is going down for reboot NOW!
What!?!?! Reboot?!? I didn't tell the script it could reboot my host. Well it did anyway. The first time we ran the script we hadn't moved all of the guests off the host and it just rebooted. Luckily they were mostly test and development guests and production wasn't impacted.
The Dell support technician maybe thought there was an issue with version 1.9 as it was just recently released, so we tried running version 1.8.1 Build 100 and had the same issue - the system rebooted in the middle of the DSET
The moral of the story - VMotion or shutdown all your guests before you run any Dell DSET tools on your ESX service console, AND, don't run Dell DSET on your ESX hosts unless you want an unceremonious reboot.
I'll keep you updated if we find out anything else.
Enjoy,
Flux.
