Dell PERC 4 (PowerEdge 2850): Unresolved configuration mismatch between disks(s) and NVRAM
Sounds like it's gone slightly nuts and is getting one config from the good drives and another slightly out of sync or invalid config from the "faulty" drive.
First thing I would do is just remove the faulty drive and try booting up then. If that doesn't work, then try with one of your other 2850's.
The problem was - as it stated in the screenshot - that no drives were detected. This was, obviously, quite a surprise as the drives were in the machine.
My hypothesis is this: when the configuration menu for the PERC was entered, it saw no disks - so rather than ask to choose between Disk Configuration and NVRAM Configuration, it showed the only configuration it knew. This presented the appearance of having checked the disks on the system when in fact, no such process had taken place - the disks remained unknown to the controller.
I also hypothesize that since the system could not access the drives electronically, it also could not detect the (known) bad drive in the array. Thus, the LED remained at a "good" state instead of bad.
I moved the disks to a new system this way:
- Shut off old system
- Shut off replacement system
- Marked all drives in the old system with their chassis slot numbers - and marked bad drive with a red tag instead of white.
- Marked all drives in the replacement system with their chassis slot numbers (just in case)
- Removed all drives from replacement system
- Rebooted replacement system and cleared configuration from PERC configuration menu
- Shut off replacement system
- Removed all drives from old system
- Placed all drives into replacement system (into matching locations)
- Rebooted replacement system
- Disabled alarm in PERC configuration menu
- Rebooted replacement system
(I also had to change the network connections, but that is not relevant to the discussion here.)
There were no problems at all going from a PERC 4e/Di to a PERC 4/DC: all of the descriptions in the manuals suggested that the only move that would not work is moving to a PERC 2 from something more recent.
After this, the system (VMware ESXi in this case) came up. There's more to do but it's all about virtual machines and VMware ESXi. The box is good.
If all remains stable, then I'll replace the bad drive with one of the ones from the retired system.
Edited for completeness
On another identical PowerEdge 2850 (also with a PERC 4e/Di), the same message came up:
1 Logical Drive(s) found on the host adapter.
0 Physical Drive(s) found on the host adapter.
After the message, no errors and the machine started normally. Thus, this message is not indicative of a failure; perhaps it only counts physical drives not in a logical drive.