Forcing a full file read in ZFS even when checksum error encountered

[Artem Belevich]

On Wed, Jul 13, 2011 at 3:01 AM, Steven Hartland
<killing at multiplay.co.uk> wrote:
> ----- Original Message ----- From: "Artem Belevich" <art at freebsd.org>
>>
>> Maybe. See here:
>> http://blogs.oracle.com/relling/entry/holy_smokes_a_holey_file
>>
>> Using that method you still would not get the bad block, but you may
>> be able to recover data beyond it.
>>
>> If you really want to get access to the data that ZFS considers to be
>> corrupted, you need to get close and personal with zdb.
>> This blog post may be a good starting point:
>> http://cuddletech.com/blog/?p=407
>
> Perfect, thanks Artem.
>
> For the record using the following zdb commands I've managed to get the
> file data off the pool.
>
> 1. zdb -dddd <pool>/<dataset> (find the objectid by searching for the
> filename)

If I understand it correctly, opjectid is used as inode number. "ls
-i" would let you find it a bit faster.

--Artem

> 2. zdb -ddddd <pool>/<dataset> <objectid> (find the indirect blocks, the
> third block of hex)
> 3. zdb -R <pool>:<indirect blocks> 2>/tmp/filedata
>
> e.g.
> zdb -dddd tank/usr
> zdb -ddddd tank/usr 1243
> zdb -R tank:0:22df120000:18000:r 2>corrupt.jpg
>
>   Regards
>   Steve
>
>
>
> ================================================
> This e.mail is private and confidential between Multiplay (UK) Ltd. and the
> person or entity to whom it is addressed. In the event of misdirection, the
> recipient is prohibited from using, copying, printing or otherwise
> disseminating it or any information contained in it.
> In the event of misdirection, illegible or incomplete transmission please
> telephone +44 845 868 1337
> or return the E.mail to postmaster at multiplay.co.uk.
>
>