Schorschi,
Regarding out-of-band/in-band flashing of iLO...
You may have misunderstood the method we used to update the iLO firmware. We DID NOT update from within ESX, but instead from the iLO interface. I am in total agreement with you in that thinking that the system console will act as a native OS to the hardware is foolish.
Saturnous,
Regarding my claim on the HP Agents crashing my blades...
Correcting my earlier statement of "crash", the actual symptom was that the system froze. It became unresponsive and after the iLO upgrade/reset I logged in through iLO and could not interact with the console even though the text was displayed. But I could not ALT-Fx and type. So I had to reset the blade.
Unfortunately, no dump files were produced and nothing was logged to the HP IML. Logically, the flashing of the iLO firmware triggered the lock, I don't think anyone can argue that - 2 systems at nearly the same time performing the same upgrade. IPMI and ILO are separate remote server management interfaces. So IPMI does not look at iLO, as iLO is not even a standard (not to mention it's a separate remote management tool). So logically, the only software piece that would have affected ESX would have been the HP Managment agents. Other than that, it was just a hardware bug, be even then iLO is suppose to stand alone.
-C
