0

We have a few X7DBT Supermicro motherboards that is using a Kingston DataTraveler 102 USB Flash Drive as a boot drive. The customer had 10 systems and they say that the site had a power failure. Once the power came back on, the 6 systems did not bootup (i.e hangs at grub stage 1.5, or cannot load fs). Grub 0.97 starts running in all cases. This is a headless system, so they shipped it back to us.

We examined the systems are our lab and found that these USB flash had filesystem corruption. I understand that flash drives like the ones we are using are unreliable, and have this paper summarizes it well.

The above state is a standard case of fs corruption, but the following reasons make this failure strange to me:

  • The drive has multiple partitions. However, none of this is ever mounted (either read only or read/write). The fs is ext2, and grub reads and a compressed fs into ram and runs from there. So, how the heck does a corruption happen if it is not mounted?

  • When we plug a couple (did not test all usb drives) into a system with X8DTT motherboard it boots fine. Both motherboards have latest bios and downloaded from Supermicro. We are going through the "prove-you-are-not-an-idiot" stage in getting support from Supermicro.

The 2nd point can be chalked up to some screwy hardware issues (or bios), but the first one is troubling. Since we are not able to recreate the problem, we need to understand this problem a little better..

Trewq
  • 163
  • 5
  • I would say the power failure did damaged the data contained in the ROM, by generating an over-voltage. – Anarko_Bizounours Aug 09 '12 at 16:07
  • If the ROM (btw, it is normal flash drive that has not been mounted) did get damaged, then it should not boot up on the other system. This system has been tested for brownouts, overvoltage, dips etc., but this of course does not mean that the system is immune. – Trewq Aug 09 '12 at 16:18
  • Well, let me explain myself in a better way. It's possible that your USB drive (even being unmounted the drive is powered) take over-voltage for a really short time (ms range), and that over-voltage was enough to modified the state inside some transistor on the ROM, so that result off corrupted data. Over-voltage don't necesseraly mean 'destruction'. Regards ;) – Anarko_Bizounours Aug 10 '12 at 14:10

1 Answers1

0

It's possible that the motherboard on the system or some other hardware component on the board was damaged and it is totally unrelated to the USB drive, the USB drive may just be collateral damage.

Regardless, there isn't much other technical data in your question to help us give you any other ideas.

Brent Pabst
  • 6,059
  • 2
  • 23
  • 36
  • Brent - If I use a different flash or reflash the usb, the motherboard works fine, so we can rule out any damage to the motherboard. The problem points to the usb flash. I have been looking at this on-and-off for a few days, but I am not sure what technical details will help. The fact it is that file corruption happens, and I do not know what caused it. I am running some 3 day tests where the machine is rebooted after certain times. – Trewq Aug 09 '12 at 17:08