Recovering from hardware failure

[pulse00]

I've been running a Freenas 10 box with 8 3TB hdds since about 2 months now without any problems. The pool i created for the 8 disks is a RAIDZ2 pool - so i should have 2 parity disks if i got it right.

Now since i came back from work today my Freenas didn't respons - neither via ssh nor the UI. I've hard-restarted the server and realized that the system keeps rebooting shortly after powering on (2-3 seconds).

There's obviously a hardware failure going on, and i'd like to ask what's the safest way to proceed with recovering from the failure.

Should i reconnect all disks from the motherboard and see if it boots without any disks attached - or should i rather disconnect one by one?

Any help would be greatly appreciated.

 

[brando56894]

Do you have a FreeNAS Mini or does your server utilize an AsRock C2750-D4I? If so the system is toast, due to a bug in the firmware that causes the watchdog to wear out one of the chips.

List your hardware either way, we have no idea what you're working with.

 

[pulse00]

It's an ASRock C2550D4I with this RAM: Crucial Ballistix Sport DIMM Kit 16GB, DDR3-1600, CL9-9-9-24 (BLS2CP8G3D1609DS1S00)

 

[ddimick]

It could be the Intel C2000 clock bug of death.

 

[Ericloewe]

Moved to hardware, since this has nothing to do with FreeNAS 10.

C2x50D4I boards have been rather unreliable. The BMC bug, which was ASRock's fault, should be fixed with the latest IPMI update.

The clock bug is not user-fixable.

Realistically, the board needs an RMA.

Also...

Crucial Ballistix Sport DIMM Kit 16GB, DDR3-1600, CL9-9-9-24 (BLS2CP8G3D1609DS1S00)

Why on earth would you not use ECC RAM? The price difference is tiny.

 

[pulse00]

If i need to change the board, what's the best way to recover from this? If i replace it with the exact same board, should i be able to recover without a reinstall?

Also, to test if it's the board: Can i disconnect all harddrives and just start the system without harddrives connected? Do i run into the risk of a failing zfs pool when doing so?

 

[Ericloewe]

If i need to change the board, what's the best way to recover from this?

There's nothing to recover from. Just replace the board.

Can i disconnect all harddrives and just start the system without harddrives connected?

I guess.

Do i run into the risk of a failing zfs pool when doing so?

If the pool is disconnected, nothing is going to happen to it...