How to recover dead FreeNAS Mini?

[Bob Paddock]

I have a ixSystems FreeNAS Mini.
After a storm that knocked out power long enough to kill the UPS
the box has never come back up.
There is no sign of life on the normal Ethernet port, VGA port or Serial Port.

There is life on both IPMI ports. Trying to do a power cycle from there responds with:

"Performing soft power off failed. This may be because the host OS is taking unusual long time to gracefully shut down."

Suggestions?

 

[DrKK]

I've never quite heard of this. Was there any type of voltage spike?

Possible failure of the SATA DOM?

Is your motherboard one of the ones affected by the ASRock IPMI-constant-reflash bug?

 

[Bob Paddock]

I've never quite heard of this.

Welcome to my life. I never seem to get 'normal' problems. :-/

Was there any type of voltage spike?

Not that I'm aware of, no other equipment on that UPS failed.

Possible failure of the SATA DOM?

That is what I suspect. Any way to test?

Is your motherboard one of the ones affected by the ASRock IPMI-constant-reflash bug?

I'm not sure, this issue?: https://www.ixsystems.com/blog/library/asrock-rack-c2750d4i-bmc-watchdog-issue/
It is a second generation style with the USB ports behind the door.

IPMI Dashboard reports:

Device Information
Firmware Revision:0.23.0
Firmware Build Time:Nov 19 2014 16:16:53 CST

Do I dare attempt the Firmware Upgrade in this state?

 

[DrKK]

I never owned an ASRock server board or FreeNAS mini. So all I know is that some of the boards have IPMI/BMC issues vis-a-vis the watchdog. I am sure you can find 10 million posts on the subject in the forum.

As for the SATA DOM, you could, I suppose, simply pull it out of the SATA port. Then, when you boot, you should get some kind of "unbootable medium" error. If everything else looks normal, then this would seem to suggest a busted SATA DOM. Then, you could try to install the matching FreeNAS version (and you can get older versions if you need) on another boot device, whether or not it's a USB drive, or another DOM, and see if you can boot (once you change the appropriate boot settings in the BIOS).

The most common failed part in systems of course is certainly not the DOM. It's the power supply. Is there any way you can just try a known-good power supply in there first?

 

[Redcoat]

I assume that you are using the Asrock utility to access the BMC/IPMI - the "Power Control and Status" page to attempt the power cycle.

My running system has a green message "Host is currently on" immediately above the first radio button line "Reset Server".

Does yours show that even though, while power is obviously supplied to the mobo, the host is clearly "not on". If that's the case it would suggest that some bits got flipped somewhere I would guess.

I just powered mine down - now I have a red message "Host is currently off". The only "available" radio button is "Power on server". "Power Cycle Server" button is inactive.

For more information on the failure mode of Dr KK's comment, look at threads here:

https://forums.freenas.org/index.ph...2550d4i-or-c2750d4i-lasted.45445/#post-307647
and here:
https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/

 

[Redcoat]

I guess that I would open it up and reseat all connectors, memory, etc., pull the battery and short the connections to clear the CMOS, before I concluded that I had a board-level problem.

 

[Bob Paddock]

I assume that you are using the Asrock utility to access the BMC/IPMI - the "Power Control and Status" page to attempt the power cycle.

Seeing that from the web interface on the IPMI port.

My running system has a green message "Host is currently on" immediately above the first radio button line "Reset Server".

Does yours show that even though, while power is obviously supplied to the mobo, the host is clearly "not on".

Yes, it says the host is on in green.

It is showing zero for +3V, +5V, +12V.
Is that normal?
All the others look normal.

https://forums.freenas.org/index.ph...2550d4i-or-c2750d4i-lasted.45445/#post-307647
and here:
https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/

Code:

inotifywait -mr -e close_write /conf/

In a bad system, you'll see the following file written to every second:
Code:
/conf/BMC1/ CLOSE_WRITE,CLOSE IPMIConfig.dat

I am not seeing a one second write, not seeing anything at all when I SSH in.
I can get in, just not seeing that one second write. Not sure this applies if
the box is already died?

 

[Redcoat]

Looks like your power supply's died:

 

[DrKK]

Dead power supply, as I said, is the most common thing. And, as it happens, the cheapest to fix. :)