[kwlug-disc] Media errors on a USB disk

Bob Jonkman bjonkman at sobac.com
Sun Oct 24 10:25:33 EDT 2010


Khalid asks:

 >Even if I can find a Windows booted PC, will the diagnostics still 
work on an ext3 filesystem?

A disk diagnostic program should just be reading the raw sectors of data. Is the sector readable?  If not, is it recoverable? If not, can the drive firmware re-map that sector to a spare?

How an OS organizes the data shouldn't make any difference to a hardware-level diagnostic program.  It shouldn't care about partition tables or partitions, file system, or empty space.

Disk diagnostics are different from  file system diagnostics, like fsck or chkdsk.  File system diagnostics can only check their own kind of file system, only within that partition.  There's no ability to use the hardware sector remapping (but the file system may support remapping), although the drive firmware's hardware sector remapping could kick in when the drive firmware detects a bad block when it's accessed by the file system diagnostic.

This is why I wonder about Spinrite.  While it seems to be an awesome program, it DOES pay attention to partition tables and partitions, so it's not just checking the raw sectors on the disk. But it checks an ext3 partition just as readily as an NTFS partition.

FWIW, I've never managed to recover a damaged drive with Spinrite.  Either the drives I checked aren't badly enough damaged, so Spinrite indicates they're good, or the drive is so badly damaged that even Spinrite can't read it.  I've had a drive that banged its heads against the backstop, run Spinrite on it for 30 days continuously, and while Spinrite marked most of the drive as "Bad" it didn't recover a thing.

--Bob.


On 10-10-23 09:52 PM, Khalid Baheyeldin wrote:
> On Sat, Oct 23, 2010 at 9:41 PM, unsolicited<unsolicited at swiz.ca>  wrote:
>
>> Khalid Baheyeldin wrote, On 10/23/2010 9:14 PM:
>>
>>   So, here is an update.
>>> I ran the e2fsck -c -y on the the disk (1TB), still in its USB enclosure.
>>>
>>> One thing I noticed is that it complained about the bad blocks inode being
>>> shared by another file. Should I be concerned about this? Or has it been
>>> fixed by fsck?
>>>
>>> Also, it took a long time. I know I can take the disk out of the enclosure
>>> (with some tinkering, this USB enclosure is not designed to be opened),
>>> but
>>> I have reasons not to for the time being.
>>>
>> Warranty, or it's a client's, presumably.
>>
> No, neither.
>
> Basically, there is no spare PC in a convenient location with power and
> network
> connection to eviscerate and connect the drive to.
>
>
>> None of your messages to date indicate you mistrust the data that's there,
>> so I'll leave that branch of this thread. Mind you, Charles' advice about
>> taking an image (now?), resonates. Presumably a raw image. Thus, if it's a
>> client's, you may be protecting your backside by taking one. (Hopefully you
>> don't have to buy another USB drive just to back up this one. See 'irony.')
>>
> The disks are used for backup, and they are not the only disks that have
> that. I do backups daily to 2 separate disks. These ones are the higher
> capacity (the other is only 500GB), and they are the fixed disks (the others
> are removable and go off site on a rotation).
>
> So all the data on these disks are replicated elsewhere and hence
> replaceable, with some moderate effort.
>
> The decision here is to detect whether these disks are no good for future
> backups, can they be trusted, what caused them to go bad (is it really
> physical platter problem, or just USB connection, or what exactly).
>
> One of them has been on constantly for maybe 1.5 years. The other was on for
> part of that.
>
>   Hopefully this is the only USB device you have hooked up at the moment, at
>> least on that channel.
>
> When running fsck, this was the only active device on USB on that machine.
> There is another on some other USB port, but it was not doing anything.
>
>
>> Any change if you use a different usb port?
>
> Not going to try it again for 14 hours.
>
>
>> The AC adapter is plugged in to it?
>
> Yes, it is powered by a separate adapter.
>
>
>> Have you noticed any difference between a 'powered' usb port and a
>> non-powered one?
>>
> Don't use any non-powered ones at the moment.
>
>
>> If warranty is the issue ... well ... how 'bad' does a drive have to be
>> (per your manufacturer) before it is replaceable? Perhaps worth finding out?
>>
> Not sure if they are under warranty or not. Need to lookup the receipts to
> know.
>
>
>> This thread does make me wonder ... are you dealing with file system
>> corruption, or a disk going bad (let alone the usb interface "isn't ATA").
>> If it's disk corruption, presumably file system corruption is not far
>> behind, jeopardizing the files that are currently there and healthy, now.
>> (You have a backup?) Could that be what you are now seeing? (Thus the inode
>> issue, now?)
>>
> The errors came up the same with the same block numbers as before. So maybe
> it is physical damage, but I can't rule out FS corruption either.
>
> Someone who knows this stuff better can chime in.
>
>
>> Since it's a USB drive and you're reluctant to extract it, I'm guessing the
>> manufacturer intended it to only ever be a USB drive. In that case ... might
>> they have a diagnostic program that does run over USB?
>
> Even if I can find a Windows booted PC, will the diagnostics still work on
> an ext3 filesystem?
>
>
> _______________________________________________
> kwlug-disc_kwlug.org mailing list
> kwlug-disc_kwlug.org at kwlug.org
> http://astoria.ccjclearline.com/mailman/listinfo/kwlug-disc_kwlug.org



More information about the kwlug-disc_kwlug.org mailing list