UBIFS Index Node Corruption - Invalid Key Type

[CN, Aananth (A.A.)]

Hi Richard,

Yes, it is the same bad key type 5.

Additional hint is, the log reported "unclean shutdown" happened. So, is a reset happened during commit operation?

Regards,
Aananth 

Sent from my mobile phone

> On 13 மே, 2016, at 7:24 பிற்பகல், Richard Weinberger <richard at nod.at> wrote:
> 
>> Am 13.05.2016 um 11:40 schrieb DHANAPAL, GNANACHANDRAN (G.):
>> Richard,
>> 
>> I believe CRC value for a node  is calculated in RAM before writing the node in NAND. 
>> In this case, node CRC and calculated CRC are matching. so Will there be possibility 
>> for corrupting key type area (32 bits) and keeping others intacted. or any other 
>> possibility for node getting corrupted.
> 
> I have to correct myself, the branches of an index node are covered by
> the node header CRC.
> This makes the whole picture more strange.
> 
> You said you're facing the same issue on another unit.
> Is the corruption really the same? Bad key type 5?
> 
> Thanks,
> //richard