I haven't had any hardware issues with power loss. Sure the file system might get messed up, but a journaling file system will help out there.
Robert
On Wed, May 12, 2010 at 5:19 PM, fred basset <> wrote:
Thanks for the response, sorry you've had to go through all those problems, sounds like a huge hassle.
As far as you know, a hard power off of the system should not cause an SD card hardware failure? (I am aware of the potential filesystem problems).
Thanks,
Fred On Wed, May 12, 2010 at 2:53 PM, Robert Ulbrich <> wrote:
I've not personally seen corruption on any of my SD cards, although my client (who is a pretty technical guy) has reported data corruption. I have seen complete hardware failure though. Like I said, I've got ~10 SD cards that I cannot get a PC to even recognize the low level device. This is after running them in a 7800 board and dumping lots of data to them in small chucks. I've run the same types of tests on a 7260 and didn't have any issues. I've also had the SD card file system on my 7800 completely lockup after a lot of use. This lockup seems to only be an issue if multiple threads are writing to the card.
For the 7800, apparently Technologic started replacing a transistor (without telling all their customers) that was under spec'd on new boards without releasing a new hardware revision. This transistor is supposed to be fine for regular SD cards (2G and less), but for SDHC cards the SD's microcontroller may become unresponsive and cause the kernel drivers to freeze (and potentially corrupt data).
Not sure if this is all related to your issue, but it might be of use.
Finally, and this is the straw that broke the camel's back, the transistor that we were told was underrated is Q9. We were told to visually inspect the transistor and look for a silver dot. Transistors that have been replaced will have the dot. None of our boards had the new transistor, so we asked Technologic to send us the transistors and we would replace them ourselves. We did so on several of our boards. Turns out that Q2 was the transistor that needed replacing. So now we have several 10s of boards which we have rendered useless, and still don't have a fix for our customer.
If you have a 7800, I would suggest calling TS and getting a refund.
Robert
On Wed, May 12, 2010 at 3:42 PM, fred basset <> wrote:
This is a 7260, which I've used before and never had this type of problem.
Could this be an actual hardware problem on the SD card itself? I'd like to know exactly how and why it happened.
On Wed, May 12, 2010 at 11:51 AM, Robert Ulbrich <> wrote:
What board are you using? I have had quite a lot of troubles using the TS-7800. I think I've killed around 10 SD cards :(.
Robert
On Wed, May 12, 2010 at 1:12 PM, fred basset <> wrote:
Has anyone else on the list seen a situation whereby an SD card
becomes corrupted?
I had an SD card running Debian in my TS system and was making some
config changes and adding some additional packages.
Over the space of a few days it seemed to be booting and running perfectly fine.
I then tried to make a backup of the SD card (2GB, Sandisk), using dd.
The dd copy would always fail with an I/O error, at the same place
in the copy each time (approx 235MB into it). I tried a number of
different PCs and SD card readers but all with the same result.
The problem has gone away with the use of a new SD card.
So what is the mechanism in play that could cause this type of error
on an SD card?
Thanks,
Fred
-- Robert Ulbrich Software Developer CAPSHER Technology, Inc. 1604 Crescent Pointe Pkwy. 979.776.7520 x150
__._,_.___ ![]()
__,_._,___
|