Failing Disk
Jim Canfield
jcanfield at tshmail.com
Thu Apr 12 14:26:28 UTC 2007
George Magklaras wrote:
> Jim, I disagree with you. I would be interested to know how dd would
> handle read errors on the failing drive. :-) Have you completed many
> rescue operations with drives whose reliability is questionable
> without hickups only with dd???
>
> If his failing drive is in a bad state and is likely to give
> persistent I/O errors, doing a dd the way you describe it in your
> number list will either abort the read operation or copy things
> inconsistently. Again I would substitute dd with dd_rescue. If his
> blocks are OK, dd_rescue will behave exactly as dd. If the blocks on
> the origin drive are broken, it will persist until it copies as much
> data as possible.
>
You are right, I mentioned previously he may have problems if the drive
was actually failing, dd_rescue never even came to mind. Thanks for
pointing it out.
-Jim
>>
>>
>> Mark,
>>
>> Did I give bad advice? I have used dd quite a bit and never had any
>> problems. Granted I am always copying to identical drives. Now
>> that I
>> think about it, it would be important to have identical disk geomerty
>> (cylinders, heads, sectors). Sorry Troy, guess I'm exposing my
>> ignorance.
>> :)
>>
>> -Jim
>
>
>
>
More information about the redhat-list
mailing list