Mailinglist Archive: opensuse (911 mails)

< Previous Next >
[opensuse] Re: "no valid journal superlock found" on mkfs & mount ext4
On Fri, 11 Dec 2015 20:23, Anton Aylward <opensuse@...> wrote:

On 12/11/2015 12:04 PM, jdd wrote:
Le 11/12/2015 17:46, Florian Gleixner a écrit :

Are you sure you don't have a fake usb stick? A common problem are usb
sticks that only have a fraction of the printed capacity.If you reach
the end of the storage device, you get redirected to the beginning.
badblks is happy with it, but filesystems are not.

if it's not a hard drive but a pen drive, test it with h2testx.exe on
windows or f3read/f3write on linux

Unhelpful.
I don't do windows and search didn't find f3read/f3write.

http://oss.digirati.com.br/f3/
(seems to have troubles, google cache has the page)

https://github.com/AltraMayor/f3
(Source on githup)


Of course I tested with dd, quite obviously, and that seemed to verify
that all 128G is writeable with no hiccups.

Bah, not listening again.

Lets say the device is fake, and has a real size of 64GB, but fakes 128GB.
The program "dd" will write the 128GB happily, it will not detect the fake.

That is what such special tools are for.
They write markers in a relative distance from each other, and check them.

Example of such tool as pseudo code:
Get device size in blocks (e.g. one per reported GB)
random_num=rnd(64bit)
Jump to start of device
Repeat loop for n blocks:
Check for marker (random_num followed by block_num)
Marker exists and reports a block_num smaller than loop_num:
Device is Fake
Abort loop, report error
Marker does not exist:
Write marker (random_num followed by loop_num)
Jump to next block start
Loop_end

dd will NOT help you here.

More links:
https://www.google.com/search?q=f3write+linuxw


- Yamaban.
< Previous Next >
Follow Ups