A common occurrence discovered by HP Proliant ML110 and ML115 owners when having installed the HP version of VMware ESXi containing the management agents is a Purple Screen of Death (PSoD) after about 2 minutes of the ESXi host being up and running. The following video demonstrates this:
The reason this occurs is that ESXi loads the HP CIM (Common Information Model) agents which then subsequently crashes ESXi if particular hardware is not found. Needless to say the hardware the CIM agents are looking for are not present in the Proliant 100 series of servers which includes the ML110 and ML115 models.
With the introduction of VMware vSphere you are now able to view the majority of hardware in your ESX(i) host and its current status via the use of the common interfaces standard, Intelligent Platform Management Interface (IPMI). Though you will notice that this doesn’t quite extend to the onboard disk controller of the Proliant ML115 (or ML110). Similar specification entry level servers from the likes of Dell do show the full disk and onboard controller details, so what’s the catch? From what I can make out HP uses their own minor variation of IPMI which talks to the BMC which then in turn talks to the ILO to gain some of the additional component information.
Herein lies part of the problem – the HP Proliant 100 series of servers don’t come with the full version of the ILO. They can have the Lights Out 100 card installed though this is the equivalent of the low fat/no frills version of the full ILO which doesn’t provide the same level of hardware detection and monitoring information.
So how do you know if one of your hard disks connected to the onboard controller is having issues? Well, the answer is – you won’t. The official line is that these servers aren’t on VMware’s hardware compatibility list (HCL) so you may want to consider is using a PCIe based array controller such as the HP Smart Array E200 or a Dell DRAC if you want this level of hardware monitoring. Both of these array controllers can often be found at a semi-reasonable price on EBay.
Unfortunately your disk controller and drive monitoring issues don’t go away if you buy an HP E200 as you will still need to load the HP Management Agents on to your ESX(i) host for the E200 to be detected and presented in the ‘Health Status’ window of ESX(i). At this point you are then back to square one as once the HP Management Agents are loaded ESX(i) will PSoD again.
To overcome this PSoD issue you will need to install the HP Management Agents and then stop all but the Smart Array controller agents from loading. Once this is complete and the ESX(i) host rebooted you should now see the HP Smart Array E200 controller and attached hard disks in the ‘Health Status’. I only have one E200 controller in my lab though have used this method since ESX 3.5 without any issues. I will put together a step by step guide if anyone is would like to see how this is done.