1

We have failures at work on our compact flash drives where the MBR gets corrupted (zeroes written to its Sector 0). The CF cards are industrial grade, but are subject to high vibration/stress environments. It has not been duplicable in our own lab even with a vibe table. I was wondering if there were "common" ways that MBRs get corrupted in order to be replicated in the lab. Thanks.

Crystal
  • 167
  • 1
  • 5
  • 1
    I think serverfault.com would give you a little more insight. It's stack overflow for sysadmins :) – Nic Jul 26 '10 at 20:22

2 Answers2

1

Most common way to corrupt a CF card is to remove it during a write. To do this with the MBR, you would need to do it while the MBR is being written. In Vista/W7, from a command prompt you can:

bootrec /FixMbr

For W2K/XP:

fixmbr \Device\HardDisk0

Those commands write to the MBR, so you need to pop the card out as soon as you run them.

JNK
  • 8,306
  • 28
  • 31
  • Would there be any difference between this happening on Windows vs Phar Lap OS? – Crystal Jul 26 '10 at 20:51
  • TBH I am unfamiliar with Phar Lap. I did some checking and couldn't find an online reference for it for the commands we need. If you have a doc, see if there is an equivalent command to rewrite the MBR and if there is you can try that. – JNK Jul 26 '10 at 20:55
0

I think there's little chance of a mechanical problem resulting in exactly a sector of 0's in sector 0.

If I had to bet, I'd bet on a software malfunction, i.e. a failure that causes the sector calculation to end up pointing at 0. Normally the opsys should keep the underlying numbers in a safe location, but something else running at kernel level (e.g. a video driver) could accidentally step on a critical location and cause funky behavior.

Alternatively, I'd consider the possibility that folks have viruses on their boxes.

Carl Smotricz
  • 648