Fusion-io ioScale issue - Internal Error

[Midorica]

Hello,

I have encountered an issue with a ioScale 3.2TB where the drive throws an error about an internal error and that the device gets disabled right after it lods the drivers.

As for the actual status, it shows the following error under fio-staus -a :

Failed: DEVICE IS OFFLINE. ALL DEADS AND WRITES WILL FAIL!



Has anyone had this issue and if so, is there any way to fix it?

I did try multiple drivers such as 3.2.15-Dell, 3.2.3-DELL and 3.2.4-DELL and there wasn't much of a difference.

Here is also a screenshot when I try to format or attach through a command, of course the same error happens when i try to detach and try the secure format.



Thanks!

 

[acquacow]

Do you have data on the drive that you need? If not, I'd move it to a proper pci-e slot and throw it thorugh a fio-sure-erase. A pcie x1 slot is not a proper home for an ioDrive. Needs to be at least pci-e 2.0 x4 for that drive.

 

[Midorica]

nah there is no data on that drive, i just plugged it and it wasn't working at all.

i did now remove one of the 3 modules plugged on the drive and now the drive show as working but running in minimal state.

I now does show up the firmware version which is 7.1.15 which i am now updating.

Ill see how it goes after the update

 

[Midorica]

Seems like that even the update, i still can't detach, attach or even format / sure-erase the drive.

The status says the following :

fct5 Status unknown: Driver is in MINIMAL MODE:
General channel initialization failure

Edit: figured out the problem but can't be really fixed... , the first nand board next to the io shield has burned marks.

 

[acquacow]

Have any closeup pics?

 

[Midorica]

I do but not on hand, ill post it in a few hours.

There was a "hole" or some sort of burnt mark next to one of the NAND chip and Windows reported that it was running at 86C under the event viewer even though there is a fan blowing air at it directly and it is cold to the touch.

Here is the photo :

 

[acquacow]

The only temp that gets reported is the FPGA itself, there's no NAND temp monitoring. I'd probably check that the heatsink is firmly attached to the FPGA with the thermal pad.

I can't really see anything obvious in the pic you posted, but if something physical happened to one of the nand boards, then yeah, that will mess some things up for sure.

I'm not sure if fio-format will let you use the remaining ones if you take one off. The NAND modules are expected to be populated from the pci-e bracket toward the FPGA. I'm not sure if a 3.2 or 6.4 would support one of the nand packs being removed though, they were never sold in partially-configured capacities like the ioDrive II half-height was.

-- Dave

 

[Midorica]

nah i tried to remove it and then try to use fio-format but it ended up in the card reporting a device failure.

In the end, it seems to be that this card is dead as my 3 other ioScale 3.2TB and the ioDrive2 Duo that i run are all working except for this one