That's a little frightening to me, that an array controller could switch to HBA only mode for no apparent reason on a Proliant server.
I gather that some P4xx models are apparently used in integrity/superdome servers in HBA only mode, but how on Earth does someone get their controller messed up like that?
Now I've got something new to worry about every time I install a firmware update (like the disastrous "CCM" flash that broke Broadcom NICs recently).
Just to inform myself, I googled some info about the P41x controllers and HBA mode... sounds like if you're using the CLI you can do a:
saupdate get_mode <controller>
For example:
saupdate get_mode all
or
saupdate get_mode p410i
Apparently that lists what mode the controllers are in, RAID or HBA.
To switch it, there's a command listed like:
saupdate set_mode <controller> <hba|raid> [-f]
For example this would change the mode for all p410i controllers to RAID:
saupdate set_mode p410i raid
(the -f will force it and not prompt the user to confirm)
Here's the kicker though... how does one send these CLI commands? The doc I'm reading mentions doing it from the UEFI shell, which you're not going to have on a Proliant.
I'm guessing there's some utility out there then that does the same esential thing? I don't see any similar option in the SSACLI, but apparently something is possible there with the right utility.
I do see that besides the UEFI partition method, the Integrity doc mentioned using an Offline Diagnostics CD. I have no idea if that CD is available online or if you can only get it if you have a valid contract on a BCS model, but it could be worth investigating. Might not boot on a Proliant even for all I know.
Hopefully HP addresses this publicly so people know what to do (like they did with that Broadcom NIC flash issue).