Re: [smartmontools-support]Re: long selftest and SAMSUNG SV1604N not working

[Bruce A. <ba...@gr...>]

Sergey,

This is VERY helpful.  Thank you!  I have one question, below.  (I am
going to write up a how-to, once I have the details sorted out).

> > so the error is at LBA = 0x016561e9 = 23421417
> > 
> > fdisk -lu /dev/hda shows:
> > /dev/hda3       5269320 238227884 116479282+  83  Linux

> First you need to calculate the sector number relative to the start of
> partition:
> 
> 23421417 - 5269320 = 18152097

Good, I agree.

> Then you need to convert the sector number to the filesystem block
> number. To do this, you need to know the filesystem block size -
> tune2fs -l /dev/hda1 will print this. Usually the block size is 4096
> bytes, which is 8 512-byte sectors.
> 
> 18152097 / 8 = 2269012.125 (the integer part gives the block number)
> So you need to do "icheck 2269012" in debugfs.

I am concerned about this.  Isn't it the case that the file system blocks
are counted from 1, not from zero?

debugfs:  icheck 0
icheck: Invalid block number 0
debugfs:  icheck 1
Block	Inode number
1	<block not found>

In this case, isn't the correct block given by 2269013 not 2269012?

Note: I do agree with the 'seek' value:

> One solution is to write zeros to the block with dd:
> 
>   dd if=/dev/zero of=/dev/hda3 bs=4096 count=1 seek=2269012


So, please let me know if you agree that the block used by debugfs needs
to be offset by one from what you have calculated.

Cheers,	
   Bruce